Repository navigation
Description
The whole open backlog of CannObserv/replicator, burned down one child at a time: live stability first, then the work sibling repos wait on, then operational conformance. Done when the host's disk is under a canonical GC, the capture consumer is live, the co-core pin tracks the current minor, and Dependabot proposes action bumps. Every other open issue is then either parked with a trigger or waits on a named external with an observable unblock condition.
- Last check-in: 2026-11-18T14:50:18Z
- Last issue seen: #211 Catch up co-core 0.48 → 0.50 (0.50.0) #213
- Last verified SHA: 753ecb9
- Next: none (parking-lot mode since 2026-11-16). Waiting: Does Replicator's Redis client send CLIENT SETINFO? (CannObserv/broker#85's Valkey 7.2 grant) #178 ← Redis 7.0 is upstream EOL (2024-07-29): decide the broker's next Redis broker#85; parked rows below
Order
The operator's rule, verbatim: "Prioritize bug fixes, issues other repos are waiting on, operational conformance and stability. Older issues may need to be re-evaluated for continued relevance."
Tiebreaks the coordinator added, numbered so the operator can challenge one:
- A hazard measured live on the host ranks as a bug. Host disk GC: adopt provisioner#61's canonical GC + headroom sensor on co-replicator (root hit 97%) #148 is the only one: root at 81% and rising (see Measured state).
- Unblocked before blocked. A blocked child's interim (a hand step) can still run ahead of its turn.
- A multi-minor co-core pin move lands alone, in an isolated venv, before any behaviour that needs the new version (the Converge storage on the cluster content-addressed model #114 rule, kept from [Epic] content.fetch-policy producer moves to Archiver #135).
- A shared runtime is built once, by the child with the most sibling waits. Viewport capture: consume the capture command, render headless, store, publish the stored-capture fact #138 builds the browser runtime and Headless-browser fetch path for JS-rendered pages (re-sited from watcher#3) #63 reuses it.
- Context freshness: the child whose surface just changed goes next.
Overrides:
- Decided 2026-10-26 (operator): Tier 1 is Retention sweep can rmdir a shard the store just created, failing the write (transient; flakes test_run_dispatches_to_the_byte_path) #159 → Host disk GC: adopt provisioner#61's canonical GC + headroom sensor on co-replicator (root hit 97%) #148 → Egress guard: IPv6 forms embedding a private IPv4 address (NAT64, 6to4, ::/96) pass the deny table #156. Corrected the same day: Retention sweep can rmdir a shard the store just created, failing the write (transient; flakes test_run_dispatches_to_the_byte_path) #159's premise was "live in production", but it's live only in CI, because production runs no sweeper. The order stands, and the swap to Host disk GC: adopt provisioner#61's canonical GC + headroom sensor on co-replicator (root hit 97%) #148 first is offered.
- Decided 2026-10-26 (operator): Bump the dev chain's virtualenv (21.7.1 → 21.14.x) and filelock (3.32.2 → 4.2.x) past 26 advisories #160 filed into Tier 3.
- Decided 2026-10-27 (operator): Egress guard: add the IANA special-purpose ranges the default deny table lacks #161 and Context budget: COMMANDS.md and ENVIRONMENT.md are over the 10k per-doc budget #162, from Egress guard: IPv6 forms embedding a private IPv4 address (NAT64, 6to4, ::/96) pass the deny table #156, filed into Tier 3.
- Decided 2026-10-28 (operator): Tier 3 runs Bump the dev chain's virtualenv (21.7.1 → 21.14.x) and filelock (3.32.2 → 4.2.x) past 26 advisories #160 → Adopt Dependabot: nothing proposes crossing this repo's dependency caps #76 → Context budget: COMMANDS.md and ENVIRONMENT.md are over the 10k per-doc budget #162 → Egress guard: add the IANA special-purpose ranges the default deny table lacks #161. The smallest goes first; Adopt Dependabot: nothing proposes crossing this repo's dependency caps #76 builds on the fresh lock; the curation makes room before Egress guard: add the IANA special-purpose ranges the default deny table lacks #161 grows ENVIRONMENT.md.
- Decided 2026-10-30 (operator): Load env files by parsing, not set -a; . — the documented load line sources the repo .env #170 and Drift test: hold every documented SocratiCode cap to the hook's default, and document the capped update #171 (re-scoped) go into Tier 3 ahead of Egress guard: add the IANA special-purpose ranges the default deny table lacks #161: Load env files by parsing, not set -a; . — the documented load line sources the repo .env #170 → Drift test: hold every documented SocratiCode cap to the hook's default, and document the capped update #171 → Egress guard: add the IANA special-purpose ranges the default deny table lacks #161. Load env files by parsing, not set -a; . — the documented load line sources the repo .env #170's documented load line contradicts the env boundary, and Drift test: hold every documented SocratiCode cap to the hook's default, and document the capped update #171's docs were just curated.
- Decided 2026-10-30 (operator): Egress guard: add the IANA special-purpose ranges the default deny table lacks #161 takes all 12 ranges, with one doc clause. Catch up co-core 0.39 → 0.40 (the current minor; #63's render field ships as 0.40.1) #172 filed into Tier 2, after Egress guard: add the IANA special-purpose ranges the default deny table lacks #161. Declined: an uncapped-launch guard (Drift test: hold every documented SocratiCode cap to the hook's default, and document the capped update #171's CR 4), and filing the three upstream skills asks.
- Decided 2026-10-31 (operator): Egress guard: add the 12 IANA special-purpose ranges Replicator's #161 added (documentation, benchmarking, ORCHIDv2, SRv6 and the rest) watcher#412 filed with Egress guard: add the IANA special-purpose ranges the default deny table lacks #161's 12 rows. Fetch: a raw ssl.SSLError from a TLS read escapes _fetch's mapping and closes as handler_error after the delivery ceiling #173 filed into Tier 1, after Catch up co-core 0.39 → 0.40 (the current minor; #63's render field ships as 0.40.1) #172, because a pin move lands alone and its brief was already posted.
100:0:0:1::/64(RFC 9780) left out. - Decided 2026-11-01 (operator): Adopt the unified
shipping-workskill: symlinks and configuration, not overrides (gregoryfoster/skills#537) #174 → Fetch: a raw ssl.SSLError from a TLS read escapes _fetch's mapping and closes as handler_error after the delivery ceiling #173 (with thepyasn1advisory refresh folded in) → test_policy_integration: the replay-into-byte-path test reads build_handler's pre-#138 'policy' kwarg and is red, unseen by CI #175 → Headless-browser fetch path for JS-rendered pages (re-sited from watcher#3) #63. The coordinator had recommended Fetch: a raw ssl.SSLError from a TLS read escapes _fetch's mapping and closes as handler_error after the delivery ceiling #173 → a separatepyasn1child → test_policy_integration: the replay-into-byte-path test reads build_handler's pre-#138 'policy' kwarg and is red, unseen by CI #175 → Headless-browser fetch path for JS-rendered pages (re-sited from watcher#3) #63 → Adopt the unifiedshipping-workskill: symlinks and configuration, not overrides (gregoryfoster/skills#537) #174. - Decided 2026-11-03 (operator): the co-core 0.40.1 floor is split from Headless-browser fetch path for JS-rendered pages (re-sited from watcher#3) #63 as Raise the co-core floor to 0.40.1 so content.fetch's render field survives decoding (split from #63) #177 and ships first, alone. Also 2026-11-03: the coordinator merged Dependabot's chore: bump astral-sh/setup-uv from 10.3.2 to 10.3.3 #176.
- Decided 2026-11-03 (operator): Does Replicator's Redis client send CLIENT SETINFO? (CannObserv/broker#85's Valkey 7.2 grant) #178 is answered with "take the grant"; no Replicator change before the broker window.
- Decided 2026-11-06 (operator): post Headless-browser fetch path for JS-rendered pages (re-sited from watcher#3) #63's announcements on Pass info.registry's render mode through to content.fetch (cannobserv v0.40.1, cannobserv#602) watcher#414 and Per-source render mode (headless-browser fetch) as Archiver policy, carried to content.fetch archiver#379, file CannObserv/cannobserv#635, and file Catch up co-core 0.40 → 0.41 (BlobStore reporting twins: keep SweepSafeLocalBlobStore's retry where calls reach it) #179 (co-core 0.41). Coordinator's order: Catch up co-core 0.40 → 0.41 (BlobStore reporting twins: keep SweepSafeLocalBlobStore's retry where calls reach it) #179, then Fetch policy: retire the unordered path once nothing can publish a frame without a generation #163 trigger A.
- Decided 2026-11-06 (operator): co-replicator's
patching-hostscadence is weekly, not the cohort's monthly; the knob statescadence weekly(beb6d3b, the coordinator, on the operator's direction). Disk GC VERSION 4: update co-replicator from the vendored pin (skills#551) #164 rescoped to the disk GC'sVERSION4. Coordinator's order: Fetch policy: retire the unordered path once nothing can publish a frame without a generation #163 trigger A, then CI: bound every job and network install step with timeout-minutes (an apt hang held test for ~2.5 h) #180. - Decided 2026-11-06 (operator): nudge maintaining-host-disk: pin the disk GC's VERSION 4, and update editor-builds.md row 1 gregoryfoster/skills#551 for Disk GC VERSION 4: update co-replicator from the vendored pin (skills#551) #164, and file CannObserv/cannobserv#640 (co-core's stale
FetchPolicyStatedocstring, from Fetch policy: retire the unordered path once nothing can publish a frame without a generation #163 A). Coordinator's order: CI: bound every job and network install step with timeout-minutes (an apt hang held test for ~2.5 h) #180, the last unblocked child; then the hand-off is the blockers. - Decided 2026-11-07 (operator): file Tests: fold tests/test_ci.py into tests/ci/ (CR 7 of #180) #181 (CR 7 of CI: bound every job and network install step with timeout-minutes (an apt hang held test for ~2.5 h) #180); not the CI wheel-cache follow-up, which the co-pypi index move (Dependabot: add the uv ecosystem once the private wheelhouse has an index it can resolve #169) would make moot. Coordinator's order: Tests: fold tests/test_ci.py into tests/ci/ (CR 7 of #180) #181. Dependabot: add the uv ecosystem once the private wheelhouse has an index it can resolve #169 part 1's un-park is proposed to the operator (Parked).
- Decided 2026-11-08 (operator): file Tests: conftest's PRODUCTION_ENV doesn't scrub REPLICATOR_CAPTURE_ENABLED or REPLICATOR_BROWSER_FETCH_ENABLED, so pytest after load_env fails 5 tests #182 (the scrub bug), next; split Dependabot: add the uv ecosystem once the private wheelhouse has an index it can resolve #169 and file its part 1 as Resolve co-core from the co-pypi Artifact Registry index, retiring the GCS wheelhouse (#169 part 1) #183, after it. Coordinator's order: Tests: conftest's PRODUCTION_ENV doesn't scrub REPLICATOR_CAPTURE_ENABLED or REPLICATOR_BROWSER_FETCH_ENABLED, so pytest after load_env fails 5 tests #182 → Resolve co-core from the co-pypi Artifact Registry index, retiring the GCS wheelhouse (#169 part 1) #183.
- Decided 2026-11-08 (operator): Disk GC VERSION 4: update co-replicator from the vendored pin (skills#551) #164, unblocked when maintaining-host-disk: pin the disk GC's VERSION 4, and update editor-builds.md row 1 gregoryfoster/skills#551 was vendored, goes before Resolve co-core from the co-pypi Artifact Registry index, retiring the GCS wheelhouse (#169 part 1) #183. Coordinator's order: Disk GC VERSION 4: update co-replicator from the vendored pin (skills#551) #164 → Resolve co-core from the co-pypi Artifact Registry index, retiring the GCS wheelhouse (#169 part 1) #183; the weekly patching run follows Disk GC VERSION 4: update co-replicator from the vendored pin (skills#551) #164.
- Decided 2026-11-09 (operator): file Catch up co-core 0.41 → 0.44 (three minors on the co-pypi index; widened from 0.42 at review) #184 (co-core 0.42), after Resolve co-core from the co-pypi Artifact Registry index, retiring the GCS wheelhouse (#169 part 1) #183: the index move keeps 0.41.0, one change per deploy. Coordinator's order: Resolve co-core from the co-pypi Artifact Registry index, retiring the GCS wheelhouse (#169 part 1) #183 → Catch up co-core 0.41 → 0.44 (three minors on the co-pypi index; widened from 0.42 at review) #184; the weekly patching run is unblocked by Disk GC VERSION 4: update co-replicator from the vendored pin (skills#551) #164 and can run between children.
- Decided 2026-11-10 (operator): file Tests: the browser sandbox test samples a renderer before it engages seccomp, and a cold launch outruns its 1.8 s poll (browser CI red on b5d4f2f) #186 (the browser sandbox test flake) and run it before Resolve co-core from the co-pypi Artifact Registry index, retiring the GCS wheelhouse (#169 part 1) #183. Coordinator's order: Tests: the browser sandbox test samples a renderer before it engages seccomp, and a cold launch outruns its 1.8 s poll (browser CI red on b5d4f2f) #186 → Resolve co-core from the co-pypi Artifact Registry index, retiring the GCS wheelhouse (#169 part 1) #183 → Catch up co-core 0.41 → 0.44 (three minors on the co-pypi index; widened from 0.42 at review) #184; docs: retire co-gcs-blob-reader as the co-gcs-blobs reader (watcher#420) #185 after Delete the co-gcs-blob-reader service account and its keys (after 2026-11-16 soak) watcher#420.
- Decided 2026-11-10 (operator): test_inflight's env-file test fails in a loaded shell and prints the broker URL with its password #187 (the broker-URL leak in a loaded-shell test) before Catch up co-core 0.41 → 0.44 (three minors on the co-pypi index; widened from 0.42 at review) #184. Coordinator's order: test_inflight's env-file test fails in a loaded shell and prints the broker URL with its password #187 → Catch up co-core 0.41 → 0.44 (three minors on the co-pypi index; widened from 0.42 at review) #184; docs: retire co-gcs-blob-reader as the co-gcs-blobs reader (watcher#420) #185 after Delete the co-gcs-blob-reader service account and its keys (after 2026-11-16 soak) watcher#420.
- Decided 2026-11-10 (operator): fold the capture contract's stale Status line (Archiver's nit on Viewport capture: consume the capture command, render headless, store, publish the stored-capture fact #138) into Catch up co-core 0.41 → 0.44 (three minors on the co-pypi index; widened from 0.42 at review) #184; no new issue, no reply on Viewport capture: consume the capture command, render headless, store, publish the stored-capture fact #138 yet.
- Decided 2026-11-11 (operator): Store viewport captures in their own bucket, co-gcs-captures (archiver#380 D4) #189 end to end (design, provisioning, build, restart) next, then Catch up co-core 0.41 → 0.44 (three minors on the co-pypi index; widened from 0.42 at review) #184, then Fetch policy tail logs no recovery after a failed cycle #188. The capture contract's Status line stays folded into Catch up co-core 0.41 → 0.44 (three minors on the co-pypi index; widened from 0.42 at review) #184; Store viewport captures in their own bucket, co-gcs-captures (archiver#380 D4) #189 rewrites the same contract, so whichever lands second rebases.
- 2026-11-12: the 2026-11-10 fold-in into Catch up co-core 0.41 → 0.44 (three minors on the co-pypi index; widened from 0.42 at review) #184 is void; Store viewport captures in their own bucket, co-gcs-captures (archiver#380 D4) #189 rewrote the capture contract's Status line. Order stands: Catch up co-core 0.41 → 0.44 (three minors on the co-pypi index; widened from 0.42 at review) #184 → Fetch policy tail logs no recovery after a failed cycle #188.
- 2026-11-13: Catch up co-core 0.41 → 0.44 (three minors on the co-pypi index; widened from 0.42 at review) #184 shipped as 0.41 → 0.44 (one pin move, one restart; tiebreak 3 holds). Next: Fetch policy tail logs no recovery after a failed cycle #188; then docs: retire co-gcs-blob-reader as the co-gcs-blobs reader (watcher#420) #185 once Delete the co-gcs-blob-reader service account and its keys (after 2026-11-16 soak) watcher#420 lands (no earlier than 2026-11-16).
- Decided 2026-11-13 (operator): file Catch up co-core 0.44 → 0.45 (co-pypi index; closure diff version-only) #192 (co-core 0.45). Coordinator's order: Fetch policy tail logs no recovery after a failed cycle #188 → Catch up co-core 0.44 → 0.45 (co-pypi index; closure diff version-only) #192; docs: retire co-gcs-blob-reader as the co-gcs-blobs reader (watcher#420) #185 after Delete the co-gcs-blob-reader service account and its keys (after 2026-11-16 soak) watcher#420 (no earlier than 2026-11-16).
- 2026-11-13: Fetch policy tail logs no recovery after a failed cycle #188 shipped (
762fd1a). Next: Catch up co-core 0.44 → 0.45 (co-pypi index; closure diff version-only) #192; then docs: retire co-gcs-blob-reader as the co-gcs-blobs reader (watcher#420) #185 once Delete the co-gcs-blob-reader service account and its keys (after 2026-11-16 soak) watcher#420 lands (no earlier than 2026-11-16). Decided 2026-11-13 (operator): file the recovery-log gap on Consumer loop logs no recovery after a failure run (backoff resets silently) processor#134 and Bus consume loops log no recovery after a failure run (fetch_facts, registry_reconcile) watcher#436; out of epic, no wait. - 2026-11-13: Catch up co-core 0.44 → 0.45 (co-pypi index; closure diff version-only) #192 shipped (
da5a19d). Next: Raise the uv_build bound past <0.13: every uv sync warns the backend excludes the running uv #195, the last unblocked child; then docs: retire co-gcs-blob-reader as the co-gcs-blobs reader (watcher#420) #185 once Delete the co-gcs-blob-reader service account and its keys (after 2026-11-16 soak) watcher#420 lands (no earlier than 2026-11-16). - 2026-11-13: Raise the uv_build bound past <0.13: every uv sync warns the backend excludes the running uv #195 shipped (
d5926ab, no restart). No child is unblocked. The hand-off is the blockers: docs: retire co-gcs-blob-reader as the co-gcs-blobs reader (watcher#420) #185 ← Delete the co-gcs-blob-reader service account and its keys (after 2026-11-16 soak) watcher#420 (no earlier than 2026-11-16); Run replicator's unit from a release, not/home/exedev/replicator(CannObserv/broker#22's rollout, replicator first) #151 ← deploying-releases: one vendored release-deploy core (fresh install and retrofit); init-project-fastapi delegates to it gregoryfoster/skills#372; Does Replicator's Redis client send CLIENT SETINFO? (CannObserv/broker#85's Valkey 7.2 grant) #178 ← Redis 7.0 is upstream EOL (2024-07-29): decide the broker's next Redis broker#85's window; Dependabot: add the uv ecosystem once the private wheelhouse has an index it can resolve #169 ← CannObserv/cannobserv#615 (parked); Fetch policy: retire the unordered path once nothing can publish a frame without a generation #163 B (parked, CannObserv/cannobserv#640 met). The weekly patching run (Standing checks) is the operator-paced work meanwhile. - Decided 2026-11-13 (operator): close Archiver patch run tonight 2026-11-13 (~22:00-23:59Z): ~1-2 min API outage at the reboot (archiver#375) #197 (archiver's patch-window notice; no impact) and post the caller correction on Adopt patching-hosts: commit the knob, then take the first monthly run under the skill (gregoryfoster/skills#313) archiver#375.
- Decided 2026-11-14 (operator): file Catch up co-core 0.45 → 0.46 (co-pypi index; closure diff version-only) #199 (co-core 0.46) and run it before Run replicator's unit from a release, not
/home/exedev/replicator(CannObserv/broker#22's rollout, replicator first) #151, as recommended. Run replicator's unit from a release, not/home/exedev/replicator(CannObserv/broker#22's rollout, replicator first) #151 is unblocked (deploying-releases: one vendored release-deploy core (fresh install and retrofit); init-project-fastapi delegates to it gregoryfoster/skills#372 Part A shipped); its first step is the vendor bump. - Decided 2026-11-14 (operator): keep the co-core closure diff; Run replicator's unit from a release, not
/home/exedev/replicator(CannObserv/broker#22's rollout, replicator first) #151 keeps Replicator's co-pypi wrapper, and the swap to the shared one is Adopt the shared co-pypi wrapper; move the wheel fetch to its own upstream-ready script #201, after it, with the wheel fetch rebuilt so it can later be proposed upstream. The operator had asked whetherwheelsstill earns its place, since no sibling has one; the evidence is in the Adopt the shared co-pypi wrapper; move the wheel fetch to its own upstream-ready script #201 body. - Decided 2026-11-14 (operator): grow co-replicator's disk rather than reclaim, filed as Grow co-replicator's root disk (83.7% after the release cutover; patching gate at 84.0%) #204 into Tier 1, ahead of Adopt the shared co-pypi wrapper; move the wheel fetch to its own upstream-ready script #201 and the ~2026-11-21 patching run. Run replicator.service as a dedicated user, not exedev (#151 follow-up) #202 and Drift check between deploys for the installed units and host configs (#151 follow-up) #203 (the Run replicator's unit from a release, not
/home/exedev/replicator(CannObserv/broker#22's rollout, replicator first) #151 follow-ups) go into Tier 3 after Adopt the shared co-pypi wrapper; move the wheel fetch to its own upstream-ready script #201, not yet ordered; Run replicator.service as a dedicated user, not exedev (#151 follow-up) #202 needs a decision on theload_env.shworkflow first. - Decided 2026-11-15 (operator): Run replicator.service as a dedicated user, not exedev (#151 follow-up) #202 builds the shared-group model. Coordinator's order: Run replicator.service as a dedicated user, not exedev (#151 follow-up) #202 next (a measured escalation path ranks as a hazard, tiebreak 1), then docs: retire co-gcs-blob-reader as the co-gcs-blobs reader (watcher#420) #185 when Delete the co-gcs-blob-reader service account and its keys (after 2026-11-16 soak) watcher#420 lands; Drift check between deploys for the installed units and host configs (#151 follow-up) #203's disposition is asked at the next check-in.
- Decided 2026-11-15 (operator): file the co-core pin move as Catch up co-core 0.46 → 0.48 (0.48.2) #206 and run it next, while docs: retire co-gcs-blob-reader as the co-gcs-blobs reader (watcher#420) #185 waits on Delete the co-gcs-blob-reader service account and its keys (after 2026-11-16 soak) watcher#420. Drift check between deploys for the installed units and host configs (#151 follow-up) #203's disposition (ask on deploying-releases: ship the drift check once (live lags main in code that runs), as the core ships gregoryfoster/skills#641, park) is still to be asked.
- 2026-11-16: Catch up co-core 0.46 → 0.48 (0.48.2) #206 shipped (
59fb240, 0.48.2 ×3). Next: docs: retire co-gcs-blob-reader as the co-gcs-blobs reader (watcher#420) #185, unblocked by Delete the co-gcs-blob-reader service account and its keys (after 2026-11-16 soak) watcher#420; then Drift check between deploys for the installed units and host configs (#151 follow-up) #203's disposition. Standing rule (from Catch up co-core 0.45 → 0.46 (co-pypi index; closure diff version-only) #199 and Catch up co-core 0.46 → 0.48 (0.48.2) #206): a pin move's "what changes" is read from a whole-wheel diff, not the CHANGELOG; twice its breaks shipped in none of the three wheels. - Decided 2026-11-16 (operator): park Drift check between deploys for the installed units and host configs (#151 follow-up) #203 on deploying-releases: ship the drift check once (live lags main in code that runs), as the core ships gregoryfoster/skills#641 and ask there for the installed-unit and host-config comparison (posted issuecomment-6906658975); no Replicator-only drift timer.
- 2026-11-16: docs: retire co-gcs-blob-reader as the co-gcs-blobs reader (watcher#420) #185 shipped (
d549a06, docs only, not deployed). The exit criterion is met: every criterion issue is closed and verified, and every other open issue is parked with a trigger (Drift check between deploys for the installed units and host configs (#151 follow-up) #203, Dependabot: add the uv ecosystem once the private wheelhouse has an index it can resolve #169, Fetch policy: retire the unordered path once nothing can publish a frame without a generation #163 B, Research: WebRecorder/WACZ archival capture path (re-sited from watcher#4) #64) or waits on a named external (Does Replicator's Redis client send CLIENT SETINFO? (CannObserv/broker#85's Valkey 7.2 grant) #178 ← Redis 7.0 is upstream EOL (2024-07-29): decide the broker's next Redis broker#85). No child is unblocked; close or parking-lot mode is the operator's call. - Decided 2026-11-16 (operator): parking-lot mode. The epic stays open,
Next: none. A fired trigger, a new co-core minor on the index, or a new bug becomes a child at the next check-in; the 12-hourly idle check is the watch. The weekly patching run (Standing checks) continues, next about 2026-11-21. - Decided 2026-11-17 (operator): Switch to shipping-work's pr merge mode (.skills/merge_mode) #208 filed and briefed out of the parking lot; the epic stays in parking-lot mode, and returns to
Next: noneonce it's verified. - 2026-11-17: Switch to shipping-work's pr merge mode (.skills/merge_mode) #208 verified; back to
Next: none. Work lands through PRs merged on GitHub with merge commits; a verified range is now<previous>..<merge commit>. - Decided 2026-11-17 (operator): CR 6 of Switch to shipping-work's pr merge mode (.skills/merge_mode) #208 (the floating
actions/checkout@v7incontext-cadence.yml) goes upstream, filed as curating-context: install-cadence.sh renders actions/checkout as a floating tag; render a SHA pin gregoryfoster/skills#675 (the template should render a SHA pin). Nothing changes here until it's vendored; the next re-render after that picks it up. - 2026-11-18: a co-core pin brief's "what changed" comes from a whole-wheel diff for every claim, module attribution included, not only the closure (third correction: Catch up co-core 0.45 → 0.46 (co-pypi index; closure diff version-only) #199, Catch up co-core 0.46 → 0.48 (0.48.2) #206, Catch up co-core 0.48 → 0.50 (0.50.0) #211). Back to
Next: noneafter Catch up co-core 0.48 → 0.50 (0.50.0) #211.
Boundary decisions
- Close what is done or superseded. Park ideas, even dormant ones, with a re-entry trigger. Proposed 2026-10-24.
- A Replicator output that an Archiver control-plane design is still deciding (WARC/WACZ archival captures: per-InfoItem policy, produced by Replicator (design) archiver#408 WACZ, Preserve video/audio media and platform transcripts (YouTube first) for video InfoItems, produced by Replicator (design) archiver#410 video, Default featured images for InfoItems from a capture of their source, delivered to the WP Information (design) archiver#411 featured images) becomes a child only when that design files or rescopes a Replicator issue. Proposed 2026-10-24.
Exit criterion
#148, #153, #138, #76, and the bugs #159 and #156, are closed and verified live (bugs added 2026-10-26). Every other open issue is parked with a trigger, or waits on a named external with an observable unblock condition.
Verify recipe
- Gate: CI by job (
test,lint,gcs (T4 against the test bucket), and since Viewport capture: consume the capture command, render headless, store, publish the stored-capture fact #138browser (the renderer's confinements, pinned headless shell)) on the merge SHA and onmain's tip. All four green on86c8986, and bounded since CI: bound every job and network install step with timeout-minutes (an apt hang held test for ~2.5 h) #180: an apt hang now fails its step at 15 min. Since Adopt Dependabot: nothing proposes crossing this repo's dependency caps #76 they're SHA-pinned, onubuntu-24.04, with--locked;tests/cipins that. TheDependabotcheck-runs are GitHub's security updates, not the gate; see Bump the dev chain's virtualenv (21.7.1 → 21.14.x) and filelock (3.32.2 → 4.2.x) past 26 advisories #160 and Adopt Dependabot: nothing proposes crossing this repo's dependency caps #76. - Deploy: the main checkout is on
main, clean, atorigin/main. When runtime code or a dependency changed,replicator.serviceActiveEnterTimestamp is after the merge,/run/replicator/build-idis the merge SHA, and the boot journal showsworker readyplus the fetch-policy replay with no decode warnings. The skills-refresh hook commits and pushes submodule bumps tomainon its own (faed2f2); a submodule-only commit needs no restart. - Baseline:
uv run pytestin the main checkout: 1914 passed, 1 skipped (test_deploy.py:880, this session runs under VS Code, not an exe.dev session root), 104 deselected, 2026-11-08 on86c8986[measured], identical in a clean shell and after. scripts/load_env.shminusREPLICATOR_REDIS_URLsince Tests: conftest's PRODUCTION_ENV doesn't scrub REPLICATOR_CAPTURE_ENABLED or REPLICATOR_BROWSER_FETCH_ENABLED, so pytest after load_env fails 5 tests #182 (a guard keeps everyboolsetting scrubbed). Save every run's full output to a file (> $SCRATCH/run.txt), never only its summary line: the watched flake's names were lost twice that way (under an exe.dev session root, sotest_deploy.py:880runs).-m browserinsystemd-run --user --scope -p Delegate=yeswithREPLICATOR_PLAYWRIGHT_BROWSERS_PATH=/var/lib/replicator/ms-playwright: 31 passed.-m integrationfrom anenv -ishell against a scratchredis-server --bind 127.0.0.1 --save '' --appendonly no(shut down after): 57 passed. Both 2026-11-06 on74c69a7[measured]. - By deliverable:
- co-core pin →
uv.lock,uv pip showfor all three packages in the service venv, and the closure-diff report in the pin comment. - new consumer stream → broker reads the group back as
acladmin(Go-live check for content.capture: confirm replicator.capture exists before archiver's first command broker#116), a first outcome in the journal, andACL LOGclean (broker reads it). Never probe the group asreplicator. - host disk →
df -h /,/var/lib/cannobserv-disk-gc/state.jsonexit0, the GC unit'sremoved … 0 failedline, andcannobserv-disk-headroomloggingchecked in ok (ok): HTTP 202(anok:line alone proves nothing). - a
deploy/file → the installed copy matches it (cpafter every edit). - docs →
tests/test_boundaries.pygreen,measure-context.sh --exactunder budget with no dead links. It rewrites.skills/context-token-counts; restore that file afterwards so the main checkout stays clean. Nothing is over since Context budget: COMMANDS.md and ENVIRONMENT.md are over the 10k per-doc budget #162, but 11 linked docs sit above 9000 (Measured state); a child that grows one past 10000 relocates in the same change. A manifestdescriptionchange is re-embedded by the session server's watcher, or by the capped driverupdatewhen no session holds the lock. - the capture path →
worker readyreadscapture: enabled; thebrowser/cgroup'smemory.max,memory.swap.max,memory.oom.groupandpids.max;worker/memory.lowis 134217728;-m browser21/21 on the host after any runtime change. - the egress guard → a new form: one test per form on the literal and the resolved path, on both fetch's transport and the browser's SOCKS guard. A new row: the literal and the resolved path on
DestinationGuard, which both transports build fromblocked_networks(...)(main.py:120,:161) [traced, Egress guard: add the IANA special-purpose ranges the default deny table lacks #161]. Live probes go in the service venv, literals only; an ISATAP or other carrier probe needs a public prefix (2600:1f18::), since2001:db8::/32is now a row. - browser fetch →
worker readyreadsbrowser_fetch: enabled; the first live one is astored a blob …line with"render": "browser";-m browserin a delegated scope. - a comment-only change in
src/→ no restart;build-idmay trailmain(first case: Fetch policy: retire the unordered path once nothing can publish a frame without a generation #163 A).
- co-core pin →
Tier 1 — live stability
- Retention sweep can rmdir a shard the store just created, failing the write (transient; flakes test_run_dispatches_to_the_byte_path) #159 Retention sweep can rmdir a shard the store just created · S · bug · live in CI (a ~1-in-50 flake of
test_run_dispatches_to_the_byte_path); latent in production, where the object-store backend starts no sweeper [measured 2026-10-26] · brief on the issue 2026-10-26 · a write-side retry is weighed against an mtime grace alone; a co-core change is draft-only · verified 2026-10-26:9b37afe..66321b5(5 commits, fast-forward, no PR); CI green by job on66321b5(main's tip); live since 18:55:53Z, build66321b5, object store, no sweeper, 0 WARNING/ERROR; four forced reproductions red on9b37afe; the grace alone was shown not to suffice, so a one-shot write-side retry was added; closed. The co-core draft is unfiled (operator's call); limits under Accepted limitations - Host disk GC: adopt provisioner#61's canonical GC + headroom sensor on co-replicator (root hit 97%) #148 Host disk GC: adopt the canonical GC and headroom sensor (CannObserv/provisioner#61) · M · dependency met: New skill: maintaining-host-disk — adopt Provisioner's canonical disk GC + headroom sensor, plus the repo-side knobs (cohort-wide) gregoryfoster/skills#425 closed 2026-10-26 (
maintaining-host-disk, GCVERSION2) · sensor via Tenant, tailnet edge + dead-man monitor for co-replicator's root disk (co-replicator / co-replicator-disk-headroom) status#67 (closed 2026-10-27) · brief revised 2026-10-27 after the Status agreement · verified 2026-10-27:66321b5..6010a66(faed2f2the hook's submodule bump, then 4 docs commits), CI green by job on both;VERSION2 installed, MANIFEST verifies; both timers enabled (split: nothing to retire, and a disabled monitor still pages); first GC run 05:25Z removed 1011.2M,exit0; sensorchecked in ok (ok): HTTP 202every 10 min; monitor enabled 16:42Z; rehearsal paged and cleared, receipt confirmed by the operator; lines 85/80 (operator); closed - Egress guard: IPv6 forms embedding a private IPv4 address (NAT64, 6to4, ::/96) pass the deny table #156 Egress guard: IPv6 forms that embed a private IPv4 address pass · S–M · bug, security, latent here (no default IPv6 route [measured 2026-10-27]) · from Viewport capture: consume the capture command, render headless, store, publish the stored-capture fact #138's review · brief on the issue 2026-10-27 · verified 2026-10-27:
6010a66..6efad10(6 commits, fast-forward, no PR); CI green by job on6efad10; live since 20:32:25Z, build6efad10; probed in the service venv: each embedded private form refused, public NAT64 and 6to4 allowed and returned unchanged; shape: checked, never substituted (the implementer's correction to the brief: the SOCKS guard dials the returned address);64:ff9b:1::/48is a row; baseline 1735/1; closed. Watcher's port filed as Egress guard: IPv6 forms embedding a private IPv4 address (NAT64, 6to4, ::/96) pass the deny table watcher#402 (operator) - Fetch: a raw ssl.SSLError from a TLS read escapes _fetch's mapping and closes as handler_error after the delivery ceiling #173 Fetch: a raw
ssl.SSLErrorfrom a TLS read escapes_fetch's mapping and closes ashandler_errorafter the delivery ceiling · S · bug, latent (0SSLErrorin the journal since 2026-09-11 [measured]) · filed 2026-10-31 (operator), from co-core 0.39.4's CHANGELOG (CannObserv/cannobserv#597) [stated there];handler.py:648catches onlyhttpx.HTTPError[traced] · reproduce test-first; close as not reproduced if httpx already maps it on this lock · Decided 2026-11-01 (operator): runs after Adopt the unifiedshipping-workskill: symlinks and configuration, not overrides (gregoryfoster/skills#537) #174, and takes thepyasn10.6.4 → 0.6.5 refresh (4 advisories inuv audit, 2026-11-01) · For Fetch: a raw ssl.SSLError from a TLS read escapes _fetch's mapping and closes as handler_error after the delivery ceiling #173 notes and the added scope on the issue 2026-11-01: 0.40.0 doesn't close the gap (main.py:117-123injects its own transport);ssl.SSLErroris not in_TRANSIENT_ERRORS· lock change, so an isolated venv · restart owed · brief 2026-11-02 on the issue: the escape is traced on this lock (anyio 4.14.2 re-raises any non-EOFssl.SSLError; httpcore 1.0.9's read map doesn't wrap it); raise it from the response stream, throughbuild_fetch_client's stack; the cheap catch stays the default;pyasn10.6.5 needs nopyasn1-modulesmove; two commits, one restart · verified 2026-11-03:337623a..0a3c5af(5130627fix,01610f8pyasn1,1029e6aCR 1, fast-forward; then the hook's0a3c5af); CI green by job on1029e6aand0a3c5af; escape measured by the implementer against a forged TLS record (before the status line too, so any post-handshake read or write); arm athandler.py:660; live since 2026-11-02T21:21:22Z, build1029e6a, NRestarts 0, replay 504/4 clean, 0 WARNING/ERROR;pyasn10.6.5 in the service venv,uv auditclean, Dependabot alerts #22 docs: curate the context surface to the 6,000-token budget (wave A, pair 3) #23–#24 docs: split the content.fetch issuer contract to the per-doc budget #26 fixed, 0 open; step 3 decided no (docstring); closed - test_policy_integration: the replay-into-byte-path test reads build_handler's pre-#138 'policy' kwarg and is red, unseen by CI #175
test_policy_integration's replay-into-byte-path test readsbuild_handler's pre-Viewport capture: consume the capture command, render headless, store, publish the stored-capture fact #138policykwarg, so it's red, and CI never runs it · S · bug (tests) · filed 2026-10-31 by Catch up co-core 0.39 → 0.40 (the current minor; #63's render field ships as 0.40.1) #172's implementer ·tests/worker/test_policy_integration.py:214readskwargs["policy"];main.py:1018passespacer=[measured]; CI runs onlytest_oom_integration.py(Run the broker-OOM suite in CI — it is the one integration file that needs no VM broker #83) · after Fetch: a raw ssl.SSLError from a TLS read escapes _fetch's mapping and closes as handler_error after the delivery ceiling #173 (operator, 2026-11-01); hands Headless-browser fetch path for JS-rendered pages (re-sited from watcher#3) #63 a green integration suite · no restart · brief 2026-11-03 on the issue: assert throughHostPacer.record/wait_seconds, not_policy; the optional signature guard wouldn't catch this rename (build_handlerstill acceptspolicy); takes Fetch: a raw ssl.SSLError from a TLS read escapes _fetch's mapping and closes as handler_error after the delivery ceiling #173's OOM integration flake (test_oom_integration.py:1017): traced toCLAIM_MIN_IDLE_MS = 50against a cycle's duration; asrc/change means stop and ask · verified 2026-11-03:64c92c0..81396a4(5 commits, fast-forward, tests anddocs/CONVENTIONS.mdonly); CI green by job on81396a4;-m integration57/57 twice against a scratch broker (coordinator,env -i); default 1820/1 twice; no restart (build1029e6a); step 3's guard already existed (test_main.py:961); OOM cause confirmed (about 30 ms of work plus the 20 ms first backoff), fixed test-only at production's 60:1 claim-idle ratio; CONVENTIONS 9968 exact; closed - Raise the co-core floor to 0.40.1 so content.fetch's render field survives decoding (split from #63) #177 Raise the co-core floor to 0.40.1 so
content.fetch'srendersurvives decoding · S · split from Headless-browser fetch path for JS-rendered pages (re-sited from watcher#3) #63 · Decided 2026-11-03 (operator): the pin ships first and alone · 0.40.1 ×3 in.wheelhouse/,renderin the wheel'sContentFetchCommand[measured]; pinspyproject.toml:285-287· test first throughfrom_wire; isolated venv;*.whlonly · brief 2026-11-03 on the issue · restart owed · verified 2026-11-03:81396a4..93fb513(3 commits, fast-forward; then the bot's7c15148, metrics only); CI green by job on93fb513; 1821/1; live since 21:46:17Z, build93fb513, 0.40.1 ×3 in the service venv, replay 504/4 clean, 0 WARNING/ERROR; decode test red on 0.40.0 viapoll_once; closure diff 1changedfile (changes.py); re-embed confirmed bycodebase_context_search; closed. Corrected in place: my brief's decode site, twice (it's co-core-aio'sAsyncBusConsumer.read, called atloop.py:1274) - Does Replicator's Redis client send CLIENT SETINFO? (CannObserv/broker#85's Valkey 7.2 grant) #178 Broker asks whether Replicator sends
CLIENT SETINFO([Epic] Burn-down: open backlog -> correctness and operational stability (2026-10) broker#80, the Valkey 7.2 window) · S · filed 2026-11-03 by broker's coordinator · traced: redis-py 7.4.1;build_bus_client(src/core/bus_client.py:112) passes nodriver_info, so it sends;rcliis redis-cli 7.0.15 and sends none · Decided 2026-11-03 (operator): take the grant (+client|setinfo), no change here · answered on the issue 2026-11-03 · stays open until broker's post-window notice on Redis 7.0 is upstream EOL (2024-07-29): decide the broker's next Redis broker#85, then broker'sACL LOGread-back shows noclient|setinfodenial forreplicator - Tests: conftest's PRODUCTION_ENV doesn't scrub REPLICATOR_CAPTURE_ENABLED or REPLICATOR_BROWSER_FETCH_ENABLED, so pytest after load_env fails 5 tests #182 Tests:
PRODUCTION_ENVdoesn't scrubREPLICATOR_CAPTURE_ENABLEDorREPLICATOR_BROWSER_FETCH_ENABLED· S · bug (tests) · filed 2026-11-08 (operator) · found by the coordinator: the watched "flake" (2026-11-03, 2026-11-07) ispytestafter. scripts/load_env.sh, deterministic, 5 named tests [measured] · red-first guard overSettings' enable flags · the issue body is the brief · no restart · verified 2026-11-08:83af705..86c8986(4 commits, fast-forward: the fix, then CR 1–3); the guardtest_destinations.py::test_every_switch_on_settings_is_scrubbedderives everyboolandbool | NoneSettingsfield, red on83af705with 3 missing names;PRODUCTION_ENVadds capture, browser fetch andREPLICATOR_CAPTURE_CGROUP;docs/TESTING.mdstates the rule, not a count; CI green by job on86c8986; coordinator re-runs: default 1914 + 1 skip in a clean shell and afterload_env.sh(minusREPLICATOR_REDIS_URL),tests/ci37, ruff clean; implementer's-m gcs16 and-m integration57; no restart; closed by the implementer - Tests: the browser sandbox test samples a renderer before it engages seccomp, and a cold launch outruns its 1.8 s poll (browser CI red on b5d4f2f) #186 Tests: the browser sandbox test samples a renderer before it engages seccomp, and a cold launch outruns its 1.8 s poll · S · bug (tests) · filed 2026-11-10 (operator) ·
browserred onmain's tipb5d4f2f(the hook's submodule bump; 1 in the last 40 CI runs); 3 of 30 local single-test runs failed (2Seccomp: 0, 1 cold start with no process seen) [measured] ·observe()returns at the first renderer seen and polls a fixed 18 × 0.1 s (test_browser_real.py:354-361) [traced]; the seccomp window is inferred, and step 1 proves it (a renderer that stays at 0 is a stop-and-report) · must stay red with seccomp off and with--no-sandbox· ran before Resolve co-core from the co-pypi Artifact Registry index, retiring the GCS wheelhouse (#169 part 1) #183 (operator, 2026-11-10) · the issue body was the brief · no restart · verified 2026-11-10:b5d4f2f..c74be97(4 commits, fast-forward,tests/worker/test_browser_real.pyonly); step 1 proved the window (61 runs: 44 renderers first seen atSeccomp: 0, all reached2in 5.8–54.7 ms,NSpidnested at every reading) [measured by the implementer]; the observer now polls untilrender()ends and asserts each pid's last full reading; red with--disable-seccomp-filter-sandboxand with--no-sandbox; 50/50 single runs; scope extended (operator): the 429 test's pacer runs on a stopped clock (== 45); CI green by job onc74be97(run 42857482011); coordinator re-runs: 15/15 single runs,-m browser31, default 1914 + 1 skip, ruff clean; no restart (build74c69a7); closed by the implementer. Limits under Accepted limitations - test_inflight's env-file test fails in a loaded shell and prints the broker URL with its password #187 Tests:
test_inflight's env-file test fails in a loaded shell and prints the broker URL with its password · S · bug (tests), secret exposure · filed 2026-11-10 by Resolve co-core from the co-pypi Artifact Registry index, retiring the GCS wheelhouse (#169 part 1) #183's implementer (CR 10) · the only failure in a loaded shell; unset the variable and the suite matches a clean one, 1959/2 [measured];PRODUCTION_ENVdoesn't listREPLICATOR_REDIS_URL, and live-broker tests readREPLICATOR_TEST_REDIS_URL[traced] · the value reached one agent transcript; rotation is the operator's call, broker side · ran before Catch up co-core 0.41 → 0.44 (three minors on the co-pypi index; widened from 0.42 at review) #184 (operator, 2026-11-10) · brief on the issue 2026-11-10 (issuecomment-6779515130): general scrub vs local delenv; no secret in a diff; the script contract unchanged (repo.envsets no broker URL; corrected 2026-11-10 in place on the brief: a stale shell URL at a reachable redis without the groups readsNOGROUP→0and opens the gate, so the file now wins); red first with a fake credential only · no restart · verified 2026-11-10:294e489..10d9ef3(merge10d9ef3;fb5acbd..0f617a1, plus the bot's2e1f6a6);REPLICATOR_REDIS_URLjoinedPRODUCTION_ENV;test_every_credential_setting_is_scrubbedderives everySettingsfield named like a credential; inflight tests compare the URL's parts bound to locals;scripts/inflight.pyreads its env file alone (95b9696); red on294e489with a fake URL (printedFAKE) [measured by the implementer]; CI green by job on10d9ef3(run 42916058234); coordinator: default 1961 + 2 skips in a clean shell and in a loaded one with the live URL set and full tracebacks (password in output: 0),tests/ci41, ruff clean; the knob'sinflightfrom a loaded shell prints0, exit 0; no restart (buildbc4c5ec); closed. Limits under Accepted limitations - Fetch policy tail logs no recovery after a failed cycle #188 Fetch policy tail logs no recovery after a failed cycle · S · observability · filed 2026-11-11 from the co-broker reboot (Run co-broker's outstanding OS patches (weekly runs of 2026-11-02 and 2026-11-09 missed) broker#141): the tail logged failures at 00:15:48Z and 00:15:55Z, then nothing, while all four consume loops logged
poll cycle recoveredat 00:16:04Z [measured] · mirrorloop.py:1431-1433inrun_policy_reader(policy.py:604); the body is the brief · no runtime risk; restart to take effect · ran next (operator, 2026-11-11) · brief 2026-11-13 on the issue (issuecomment-6839748883)::575resetsconsecutive_failuresafter a malformed frame, a second silent recovery path; log both · restart owed · verified 2026-11-13:6eb65b8..762fd1a(PR #188: the fetch-policy tail logs its recovery after failed cycles #193,f4298a3+ CR 2601b4df, merged--no-ff, merge tree =601b4df); one helper_recovered()logs INFOfetch policy poll cycle recoveredwithafter_failuresfrom both ends of a run (policy.py:576,:605);skipsresets stay silent; the stuck-cursor path logs an ERROR, not the WARNING (implementer's refinement of the brief, kept); 5 cases test-first, mutation-checked on both paths [stated by the implementer]; CI green by job on601b4dfand762fd1a; coordinator: live since 2026-11-13T19:00:28Z, build762fd1a, NRestarts 0, all threeenabled, replay 501/10 hosts, 0 WARNING/ERROR; default 2005 + 2 skips clean and loaded,tests/ci41, ruff clean [measured]; closed. Same gap in the siblings, filed (operator): Consumer loop logs no recovery after a failure run (backoff resets silently) processor#134, Bus consume loops log no recovery after a failure run (fetch_facts, registry_reconcile) watcher#436 - Grow co-replicator's root disk (83.7% after the release cutover; patching gate at 84.0%) #204 Grow co-replicator's root disk · S · live hazard (tiebreak 1) · filed 2026-11-14 (operator chose to grow over reclaiming) · root 83.72%, 3.02G free after Run replicator's unit from a release, not
/home/exedev/replicator(CannObserv/broker#22's rollout, replicator first) #151's two releases (248M each); the patching gate's real ceiling is 84.0%, and each ship's deploy builds a third release (~85% peak, the sensor's alert line) [measured] · exe.dev documentsresize <vm>with--cpu/--memoryonly; disk is--diskat creation (20 GB); a disk resize is a request: the operator checks the CLI's own help or asks support, never an undocumented flag [read] · coordinator recommends 30 GB (~56% steady) · the issue body is the brief · before Adopt the shared co-pypi wrapper; move the wheel fetch to its own upstream-ready script #201 and the ~2026-11-21 patching run · restart only if the grow reboots · verified 2026-11-15: the operator's grow to 30 GiB rebooted the VM (boot 2026-11-14T23:55:50Z); ext4 sits on/dev/vdawith no partition, so no root commands; thenb61703a..7706bf0(a58615e,7706bf0, docs and theKEEP=0comment;KEEPstays 0, operator), deployed byscripts/deploy.shat 00:13Z; the sysctl drop-in installed with the operator's authorization (00:30Z) [stated by the implementer]; CI green by job on7706bf0(a58615ehad no run of its own); coordinator:/dev/vda32212254720 B, root 54.46%, 12.75G free, 29.42G usable;live→releases/7706bf05cc16(previousb61703aee439), worker pid 10051 cwd the release, 0 mappings under/home,BUILD_ID=7706bf05cc16, NRestarts 0, handler silent,systemctl --failedempty, noreboot-required; replay 503/10, all threeenabled, 22 INFO only; sensorheld_sincenull, 55%, checked in ok; the drop-in's sha256 equals the repo's (6b42b277…),vm.swappiness10,vm.min_free_kbytes65536; default 2028 + 1 skip clean and loaded,tests/deploy32,tests/ci44, ruff clean [measured]; closed by the implementer. Limit: the first GC run after the grow is 2026-11-15T05:38Z, unread
Tier 2 — sibling waits
- Catch up co-core 0.37.0 → 0.39.1 ahead of the capture consumer #153 Catch up co-core 0.37.0 → 0.39.1 · S–M · unblocked · isolated venv, closure diff · prerequisite for Viewport capture: consume the capture command, render headless, store, publish the stored-capture fact #138 (capture contracts are in 0.37.1) · verified 2026-10-25:
42b728c..2790a53(PR #153 chore: catch up co-core 0.37.0 -> 0.39.1 #155), CI green by job on2790a53; deployed 2026-10-24T22:36Z, build2790a53, co-core 0.39.1 ×3, replay clean, 20 blobs since. Floor>=0.39.0,<0.40(operator, at ship). Closed - Viewport capture: consume the capture command, render headless, store, publish the stored-capture fact #138 Viewport capture: consume
content.capture, render, store, publish · L · dependency met: co-core 0.39.1 live (Catch up co-core 0.37.0 → 0.39.1 ahead of the capture consumer #153, 2026-10-24) · design doc first: a new command stream and the capture issuer contract · approval gate: the browser runtime dependency and its memory cap · Viewport screenshots: per-source capture policy, issued to Replicator on a dedicated command stream archiver#380, Default featured images for InfoItems from a capture of their source, delivered to the WP Information (design) archiver#411 and Go-live check for content.capture: confirm replicator.capture exists before archiver's first command broker#116 wait on it · broker grants live since 2026-10-20 · createreplicator.capturebefore Archiver's first command, never at$on a non-empty stream · brief on the issue 2026-10-25, with three operator decisions (runtime, enable flag, concurrency) · verified 2026-10-26:2790a53..9b37afe(PR #138 PR 1/2: the browser runtime, its SOCKS egress guard and the delegated cgroup #1576580979, PR #138 PR 2/2: the content.capture consumer, off by default #1589b37afe); CI green by job incl.browseron both; live since 2026-10-26T16:42Z, capture enabled, unit and cgroup as designed; closed. Still open: the first real capture (Viewport screenshots: per-source capture policy, issued to Replicator on a dedicated command stream archiver#380) and broker's read-back (Go-live check for content.capture: confirm replicator.capture exists before archiver's first command broker#116), see Verification debt - Headless-browser fetch path for JS-rendered pages (re-sited from watcher#3) #63 Headless-browser fetch path · L · rework: selection is settled as option 1, an issuer-driven co-core
renderfield (Per-source render mode (headless-browser fetch) as Archiver policy, carried to content.fetch archiver#379, 2026-10-14); the body's co-core 0.10.0 field list is stale · runtime met:Renderer, SOCKS guard andbrowser/cgroup from Viewport capture: consume the capture command, render headless, store, publish the stored-capture fact #138 · waited on CannObserv/cannobserv#602, therenderfield (filed 2026-10-28, operator); met 2026-11-01: v0.40.1 is on the index and in the local.wheelhouse/[measured] · For Headless-browser fetch path for JS-rendered pages (re-sited from watcher#3) #63 notes on the issue 2026-10-26 · Per-source render mode (headless-browser fetch) as Archiver policy, carried to content.fetch archiver#379 waits on it · needs co-core 0.40.1 (render), so Catch up co-core 0.39 → 0.40 (the current minor; #63's render field ships as 0.40.1) #172 goes first · Pass info.registry's render mode through to content.fetch (cannobserv v0.40.1, cannobserv#602) watcher#414 waits on its support announcement too · the 0.40.1 floor split off as Raise the co-core floor to 0.40.1 so content.fetch's render field survives decoding (split from #63) #177 (operator, 2026-11-03); this keeps a design doc (contract change), the browserFetcher(today'sRenderedcarries only a PNG and status), the two refusal rows and the announcement · after Raise the co-core floor to 0.40.1 so content.fetch's render field survives decoding (split from #63) #177 · For Headless-browser fetch path for JS-rendered pages (re-sited from watcher#3) #63 notes from Fetch: a raw ssl.SSLError from a TLS read escapes _fetch's mapping and closes as handler_error after the delivery ceiling #173 on the issue 2026-11-03: the browser path needs its own transient-or-permanent map, tested through the worker's real client; the_fetchTLS arm goes dead if co-core'sRetryAsyncTransportis ever mounted · For Headless-browser fetch path for JS-rendered pages (re-sited from watcher#3) #63 notes from test_policy_integration: the replay-into-byte-path test reads build_handler's pre-#138 'policy' kwarg and is red, unseen by CI #175 on the issue 2026-11-03: run-m integrationlocally ifrun()wiring changes; mirror kwargs in both recorders; CONVENTIONS 9968 of 10000 · the last unblocked child once Raise the co-core floor to 0.40.1 so content.fetch's render field survives decoding (split from #63) #177 ships; the announcement to Pass info.registry's render mode through to content.fetch (cannobserv v0.40.1, cannobserv#602) watcher#414 and Per-source render mode (headless-browser fetch) as Archiver policy, carried to content.fetch archiver#379 is an external reply, so it needs the operator's consent · brief 2026-11-03 on the issue: design doc first (contract change), stop for approval;Renderedcarries no document or headers; the renderer exists only while capture is on (main.py:916), so no silent HTTP fallback; serialise renders under the one 1536 MiB cgroup; reuse capture's error map; branch onrender == "browser"; the coordinator posts the Pass info.registry's render mode through to content.fetch (cannobserv v0.40.1, cannobserv#602) watcher#414 and Per-source render mode (headless-browser fetch) as Archiver policy, carried to content.fetch archiver#379 announcements after verifying, with the operator's consent · verified 2026-11-06:93fb513..0de3ce5(the bot's7c15148, then 15 commits, fast-forward: the design doca619cb2, approved507c845with all five operator calls as recommended); CI green by job on0de3ce5; coordinator re-runs: default 1902/0,-m browser31,-m integration57; live since 2026-11-05T21:09:43Z, build0de3ce5,browser_fetch: enabled(the operator's flag,root:exedev 640), replay 508/4 clean, 0 WARNING/ERROR; CR 1'sfile:read closed bycheck_browsable; outcome reference 9985 of 10000; announced on Pass info.registry's render mode through to content.fetch (cannobserv v0.40.1, cannobserv#602) watcher#414 and Per-source render mode (headless-browser fetch) as Archiver policy, carried to content.fetch archiver#379, and CannObserv/cannobserv#635 filed (operator, 2026-11-06); closed - Catch up co-core 0.40 → 0.41 (BlobStore reporting twins: keep SweepSafeLocalBlobStore's retry where calls reach it) #179 Catch up co-core 0.40 → 0.41 · S · filed 2026-11-06 (operator), from the cannobserv coordinator's epic comment (2026-11-05) · v0.41.0 ×3 on the index and in
.wheelhouse/[measured]; pinspyproject.toml:310-312· theBlobStorereporting twins:SweepSafeLocalBlobStorekeeps working, but a twin call would skip its retry; coordinator leans to a guard test ·FetchPolicyEmit.generationrequired (Emit only) · brief 2026-11-06 on the issue · watch the co-pypi index move (stop and ask) · restart owed · verified 2026-11-06:53918a7..74c69a7(5 commits, fast-forward, on the hook's53918a7); option (b) kept with a dispatch test, an override check and asrc/twin scan proved red; CI green by job on74c69a7(teston attempt 2 after an apt hang, filed as CI: bound every job and network install step with timeout-minutes (an apt hang held test for ~2.5 h) #180); coordinator re-runs: default 1907/0,-m browser31,-m integration57; live since 2026-11-06T14:39:11Z, build74c69a7, 0.41.0 ×3, replay 508/4 clean, 0 WARNING/ERROR, all threeenabled; the manifest re-embed held (watcher active, context search reads 0.41.0); the index didn't change (CannObserv/cannobserv#554 publishes to both); closed by the implementer - Catch up co-core 0.39 → 0.40 (the current minor; #63's render field ships as 0.40.1) #172 Catch up co-core 0.39 → 0.40 · S · filed 2026-10-30 (operator): v0.40.0 shipped 14:07Z and the pin is
<0.40; the changelog lists Replicator as "downstream, no change";src/names neither tightened Emit [measured] · lands alone (tiebreak 3); takes 0.40.1 if it has shipped · after Egress guard: add the IANA special-purpose ranges the default deny table lacks #161 (shipped) · restart owed · brief 2026-10-31 on the issue: the local.wheelhouse/already has 0.40.0 (the 16:00Z restart's sync), the index has no 0.40.1, so don't wait for it;co_core_aio/co_core_sync0.40.0 requireco-core<0.41,>=0.40; pins atpyproject.toml:264-266; let the closure diff judge 0.39.2's sync-transport change (GcsBlobStoreis in the closure);AGENTS.md:35names the pin · verified 2026-11-01:37b6eca..2a6e0f2(4 commits, fast-forward); CI green by job on2a6e0f2(main's tip; the redDependabotrun is a pyasn1 security job failing on Dependabot: add the uv ecosystem once the private wheelhouse has an index it can resolve #169's PyPI blocker, not the gate); live since 22:46:02Z, build2a6e0f2,uv pip show0.40.0 ×3 in the service venv, replay 500/4 clean, 0 WARNING/ERROR; lock wheels-only; 1819/0; the manifest re-embed was done by the coordinator session's watcher (22:47:26Z), not owed; closed. Floor 0.40.0: no 0.40.1 at lock time - Catch up co-core 0.41 → 0.44 (three minors on the co-pypi index; widened from 0.42 at review) #184 Catch up co-core 0.41 → 0.44 (retitled 2026-11-13 from 0.42) · S · filed 2026-11-09 (operator): v0.42.0 tagged
ce0277f52026-11-08, on the co-pypi index ×3 [measured]; breaking only in the WordPress facades, and no removed name appears insrc/,scripts/ortests/[traced] · first use of Resolve co-core from the co-pypi Artifact Registry index, retiring the GCS wheelhouse (#169 part 1) #183's index closure-diff path · For Catch up co-core 0.41 → 0.44 (three minors on the co-pypi index; widened from 0.42 at review) #184 and pre-flight on the issue 2026-11-10 (issuecomment-6779515637):co-pypi.sh wheels+--wheels; the lock gains URLs and hashes;test_dependency_index.py:63moves to>=0.43; the pin commentpyproject.toml:10-37· isolated venv · restart owed · the issue body is the brief · the 2026-11-10 fold-in (the capture contract's Status line) is void: Store viewport captures in their own bucket, co-gcs-captures (archiver#380 D4) #189 rewrote it · refreshed 2026-11-12 on6e0b19f(issuecomment-6813911960): surface untouched; read the blobstore modules ifchanged; run-m gcs· verified 2026-11-13:6e0b19f..6eb65b8(PR #184: co-core 0.41 → 0.44 (three-minor catch-up, first pin move on the index) #191, 8 commits, merged--no-ff); widened to 0.44 at review (CR 2: 0.43.0 and 0.44.0 were on the index; the coordinator's refreshes had not re-read it, corrected on the issue, issuecomment-6839748415); closure diff 0.41.0 → 0.44.0 from index wheels: 58/58, onechanged(changes.py: only the producer-sideContentFetchCommandEmit, CannObserv/cannobserv#635), blobstore surface byte-identical [measured by the implementer]; CI green by job on6eb65b8(run 43324674811); coordinator: live since 2026-11-12T23:18:59Z, build6eb65b8, NRestarts 0, 0.44.0 ×3 in the service venv,uv lock --checkclean, replay 505/10 hosts,storing captures in their own bucket, 0 WARNING/ERROR; index newest 0.44.0 ×3 [measured]; default 2001 + 2 skips clean and loaded,tests/ci41, ruff clean; closed. Index friction recorded on Resolve co-core from the co-pypi Artifact Registry index, retiring the GCS wheelhouse (#169 part 1) #183 (probe at<1) - Store viewport captures in their own bucket, co-gcs-captures (archiver#380 D4) #189 Store viewport captures in their own bucket,
co-gcs-captures(Viewport screenshots: per-source capture policy, issued to Replicator on a dedicated command stream archiver#380 D4) · M · sibling wait (Archiver ships its reader dark; no joint window, no deadline) · filed 2026-11-11 from the operator's D4 decision · both production captures are inco-gcs-blobsand were persisted intoco-gcs-replicator(16:06:54Z, 19:11:10Z) [measured]; onestoreserves all four handlers (main.py:767) [traced] · contract change: design doc first; one bucket holding two tiers conflicts on soft delete,updateand the lifecycle scope; persist routes by source; replicate starts refusing captures (invalid_source); operator provisioning before the restart · ran end to end before Catch up co-core 0.41 → 0.44 (three minors on the co-pypi index; widened from 0.42 at review) #184 (operator, 2026-11-11) · brief on the issue 2026-11-11 (issuecomment-6809453345) · restart owed · verified 2026-11-12:10d9ef3..6e0b19f(PR #189: captures in their own bucket, co-gcs-captures (archiver#380 D4) #190, merged asba37286, then6e0b19fthe provisioning record); design approved;REPLICATOR_CAPTURE_BUCKETwith three boot refusals; prefixestemp/blobsas constants; persist routes by family; replicate refuses a captureinvalid_sourceby name, the alias guard covers the bucket; a derived*_bucketscrub guard; local captures root<blob_dir>/captures/, swept against one ceiling; provisioned by the operator (two lifecycle rules eachmatchesPrefix: temp/, 7-day soft delete,updateconditioned totemp/, object-level probe PASS ×3 [measured by the implementer]); CI green by job onba37286and6e0b19f; coordinator: live since 02:41:35Z, buildba37286, NRestarts 0,storing captures in their own bucket(co-gcs-captures,temp,blobs), 0 WARNING/ERROR; first post-split capture01M76WR77JYHK5QFS31W3JXQ44kept atgs://co-gcs-captures/blobs/a0b1c585….bin; the bucket holds 1temp/and 1blobs/object, probe objects removed; default 2001 + 2 skips clean and loaded,tests/ci41, ruff clean; closed. Implementer's corrections to the brief, kept: replicate refusedblob_expiredon a local host (nowinvalid_sourcefirst); theage: 365backstop needed the prefix too. Limits under Accepted limitations - Catch up co-core 0.44 → 0.45 (co-pypi index; closure diff version-only) #192 Catch up co-core 0.44 → 0.45 · S · filed 2026-11-13 (operator), from the idle check: v0.45.0 tagged
22ecff8c2026-11-13T05:36Z, on the co-pypi index ×3 [measured, unbounded]; pin>=0.44.0,<0.45· closure diff 0.44.0 → 0.45.0 from index wheels: 58/58, 3version-only(the__init__s), Requires-Dist the floor bump only [measured by the coordinator]; CHANGELOG: co/v1EventModel.url, sdists tosdists/, workstationCLOUDSDK_CONFIG[stated] · no code change expected · the issue body is the brief · isolated venv · ran after Fetch policy tail logs no recovery after a failed cycle #188 (tiebreak 3) · pre-flight 2026-11-13 on762fd1a(issuecomment-6862388261): index newest still 0.45.0;tests/test_dependency_index.py:62-71(not in >=0.45, docstring) goes red on the lock move and is in scope · restart owed · verified 2026-11-13:762fd1a..da5a19d(PR #192 co-core 0.44 → 0.45, one-minor catch-up #194,5c56392..7de34b7, merged--no-ff); index probed at<1at pickup, nothing past 0.45.0; pin>=0.45.0,<0.46×3; closure report 0.44.0 → 0.45.0 verbatim in the pin comment (58/58, 3version-only); the implementer found two more literal sites (.socraticodecontextartifacts.json:166,docs/DEPENDENCIES.md:86), and CR 1 corrected my body'sEventModel.urlwording (falls back to theevent_url-typed link); the manifest re-embedded [stated by the implementer]; CI green by job on7de34b7andda5a19d(run 43504680079); coordinator: live since 2026-11-13T20:48:46Z, buildda5a19d, NRestarts 0, 0.45.0 ×3 in the service venv,uv lock --checkclean, replay 503/10 hosts,storing captures in their own bucket, all threeenabled, 0 WARNING/ERROR; index newest 0.45.0 ×3; default 2005 + 2 skips clean and loaded,tests/ci41, ruff clean [measured]; closed. Follow-up: Raise the uv_build bound past <0.13: every uv sync warns the backend excludes the running uv #195 - Catch up co-core 0.45 → 0.46 (co-pypi index; closure diff version-only) #199 Catch up co-core 0.45 → 0.46 · S · filed 2026-11-14 (operator), from the idle check: v0.46.0 and v0.46.1 tagged 2026-11-13T21:14Z and 22:21Z, both on the co-pypi index ×3 [measured 13:57Z, unbounded]; pin
>=0.45.0,<0.46· closure diff 0.45.0 → 0.46.1 from index wheels: 58/58, 3version-only; Requires-Dist the floor bump plus[extract]'s soupsieve>=4.1,<5, an extra Replicator doesn't declare (no soupsieve inuv.lock) [measured by the coordinator]; CHANGELOG: 0.46.0 breaks only co/v1 WordPress models (wp_view_*_urlremoved,wp_linkfilled,EventModel.urlwithout thelinkfallback) [stated] · no code change expected; pins>=0.46.0,<0.47(no 0.46.1 feature depended on) · the issue body is the brief, with every literal site fromgit grep(the manifest's:166owes acodebase_update) · isolated venv · ordered before Run replicator's unit from a release, not/home/exedev/replicator(CannObserv/broker#22's rollout, replicator first) #151 (operator, 2026-11-14; tiebreak 3) · restart owed · verified 2026-11-14:a46e650..91e87dc(PR #199 co-core 0.45 → 0.46, one-minor catch-up #200,203c25e,d6d9d39, merged--no-ff); index probed at<1at pickup, nothing past 0.46.1; pins>=0.46.0,<0.47×3, lock 0.46.1 ×3 with hashes, no soupsieve; closure report 0.45.0 → 0.46.1 verbatim in the pin comment; manifest re-embedded [stated by the implementer]; corrected in place on the issue body: my brief'sWordPressMetadata.wp_view_*_urlremoval isn't in any of the three wheels Replicator locks, at either version (the implementer's whole-wheel diff; the wheels change onlyco_v1/responses.py,co_v1/event.pyand two docstrings, all outside the closure) [measured by the implementer]; CI green by job ond6d9d39and91e87dc; coordinator: live since 2026-11-14T19:01:13Z, build91e87dc, NRestarts 0, 0.46.1 ×3 in the service venv,uv lock --checkclean, replay 509/10 hosts,co-gcs-captures, all threeenabled, 19 INFO and nothing else; index newest 0.46.1 ×3; no<0.46literal left; default 2009 + 1 skip clean and loaded,tests/ci44, ruff clean [measured]; closed by the implementer - Catch up co-core 0.46 → 0.48 (0.48.2) #206 Catch up co-core 0.46 → 0.48 (0.48.2) · S · filed 2026-11-15 (operator), from check-in 40's index read: v0.47.0, v0.48.0, v0.48.1 tagged 2026-11-14T20:16Z to 2026-11-15T04:07Z, then v0.48.2 at 06:51:22Z, all ×3 on the index [measured, unbounded]; pin
>=0.46.0,<0.47· breaking only in the WordPress disclosure and transcript models · closure diff 0.46.1 → 0.48.2 [measured]: 5 of 58 not identical, 3version-only,co_core/effects/gdrive.pyandco_core_sync/drivers/gdrive.pychanged(Drive retries anddelete_if_presentonly, read with--diff); Replicator has no Drive writer (provider_disabled) · body plus the 0.48.2 pre-flight (issuecomment-6897799238) · verified 2026-11-16 (issuecomment-6903567908):f2a4333..59fb240(3 commits + merge, PR #206: co-core 0.46 → 0.48, two-minor catch-up #207); CI green by job on59fb240and096afd3; livereleases/59fb24082f7e(previousf2a43335c6b0), pid 39128 asreplicator, confined, NRestarts 0; release venv 0.48.2 ×3;worker readyall threeenabled, replay 503/10, INFO only, handler silent; index newest still 0.48.2; coordinator 2061 + 2 skips clean and loaded and 180 + 1 host deploy tests, the extra skiptest_deploy.py:1049from a detached run, which passes in the foreground (= 2062/1, 181); ruff anduv lock --checkclean; the code index caught up (noupdateowed) [measured]; closed by the implementer
Tier 3 — operational conformance
- Adopt Dependabot: nothing proposes crossing this repo's dependency caps #76 Adopt Dependabot · S · the
github-actionshalf, in Adopt Dependabot: nothing proposes crossing this repo's dependency caps archiver#192's shape · brief 2026-10-28 · verified 2026-10-30:f3db89b..cda2861(6 commits, then Dependabot's chore: bump actions/checkout from 5.1.0 to 7.0.1 #166–chore: bump astral-sh/setup-uv from 5.4.2 to 10.3.2 #168 merged one at a time); CI green by job on3f8a90dandcda2861; SHA-pinned (checkout v7.0.1, setup-uv v10.3.2, auth v3.0.0),ubuntu-24.04,prune-cache: trueand--lockedin all four jobs, pinned bytests/ci(20); Dependabot runs mint OIDC (row 3, blocker 1 not real); closed (operator). The uv half was re-filed as Dependabot: add the uv ecosystem once the private wheelhouse has an index it can resolve #169, parked - Bump the dev chain's virtualenv (21.7.1 → 21.14.x) and filelock (3.32.2 → 4.2.x) past 26 advisories #160 Bump the dev chain's
virtualenvandfilelockpast 26 advisories · S · filed 2026-10-26 (operator) · brief 2026-10-28 · verified 2026-10-28:02aa79f..f3db89b(the hook'sf404530, thenf3db89b,uv.lockonly), CI green by job on both; service venvvirtualenv21.15.1,filelock4.2.4,python-discovery1.6.7,packaging26.3;uv audit --frozenclean (79 packages); no restart (dev-only; build-id stays02aa79f); no Dependabot security run since the merge; closed. Corrected in place: the brief's 21.14.x option and the body's 21.14.9 floor (the floor is 21.14.13) - Raise the uv_build bound past <0.13: every uv sync warns the backend excludes the running uv #195 Raise the
uv_buildbound past<0.13: everyuv syncwarns · S · operational conformance (deploy and CI noise) · filed 2026-11-13 by Catch up co-core 0.44 → 0.45 (co-pypi index; closure diff version-only) #192's implementer (operator) ·pyproject.toml:513unchanged since Set up project foundation #1; host uv 0.13.5, CI uv 0.13.8 (setup-uvtakes noversion:) [measured] · ran as the last unblocked child · the body is the brief; pre-flight 2026-11-13 onda5a19d(issuecomment-6865003153): the surface is one line, no doc states the rule; the host's uv cache holds nouv-buildafter the 20:48Z sync, which weighs against the body's [inferred] PyPI fetch; notests/ciguard onsetup-uvversion:(4 sites); restart likely not owed [traced] · isolated venv · verified 2026-11-13:da5a19d..d5926ab(PR #195: uv_build bound to >=0.13.0,<0.14, CI's uv pinned to the same range #196,47b2334..136fc5b, merged--no-ff);uv_build>=0.13.0,<0.14; CI's foursetup-uvsteps take the range verbatim asversion:, held equal bytests/ci/test_workflow_pins.pywith a bite test; the rule in the[build-system]comment anddocs/DEPENDENCIES.md(bump all five when the host's uv moves a minor);required-versionrejected (it would refuse the unit'suv runon a host uv upgrade, measured in scratch); the body's PyPI-fetch inference disproved withuv sync -vand corrected in place; restart not owed (the installed editable is byte-identical but for uv's cache stamp) [stated by the implementer]; CI green by job on136fc5bandd5926ab(run 43512090888: uv 0.13.8, 0does not contain the current uvlines; 4 onda5a19d); coordinator: hostuv sync --frozen --dry-runmakes no change, no warning; service untouched (buildda5a19d, NRestarts 0, 0 WARNING/ERROR); default 2008 + 2 skips clean and loaded,tests/ci44, ruff clean [measured]; closed - Weekly OS patching run on co-replicator, 2026-11-14 (patching-hosts, maintenance lane: 92 packages + reboot) #198 Weekly OS patching run, 2026-11-14 (standing check; maintenance lane, 92 packages + reboot) · S · operational conformance · filed 2026-11-13 by its implementer from the coordinator's prompt · verified 2026-11-14: bulk 88 packages 00:33–00:45Z, then the held steps redis, containerd and tailscale 1.104.7 one at a time, each operator-approved; inflight
0before the reboot; boot 00:50:45Z, kernel 6.12.113 → 6.12.115, 32 s down;probe.sh --post-boot12/12 ok; after-probe 0 pending, 0 findings; knob exceptionsetup-script:redelivered(owner,3d16e77, review by 2027-02-14); upstream patching-hosts: dormancy reads a masked local redis as in-use because a deploy/ unit runs a script whose filename contains 'redis' gregoryfoster/skills#634 filed, data point on patching-hosts: a dry run counted against stale lists writes count=0, and apply.sh's bulk gate accepts it in the security and maintenance lanes gregoryfoster/skills#629, process log patching-hosts process log: 2026-11-14 replicator (overdue weekly run; stale-list dry run aborted on a 404; disk-gated at 84.0%) gregoryfoster/skills#639 [stated by the implementer]; coordinator: booted 00:50:45Z on 6.12.115;worker readybuild3d16e77, all threeenabled, replay 503/10 hosts, NRestarts 0, 0 WARNING/ERROR;apt list --upgradable0; noreboot-required;systemctl --failedempty; tailscale 1.104.7,AutoUpdate.Applyfalse; headless-shellldd0 not found; both disk timers scheduled; root 80%; CI green by job on3d16e77anda46e650[measured]; closed - Run replicator's unit from a release, not
/home/exedev/replicator(CannObserv/broker#22's rollout, replicator first) #151 Run the unit from a release, not the dev checkout · L · survey done 2026-10-23 · blocked by deploying-releases: one vendored release-deploy core (fresh install and retrofit); init-project-fastapi delegates to it gregoryfoster/skills#372 (shipping-work*: a.skills/deploy_commandknob — ship through the project's deploy script instead of "migrate, then systemctl restart" gregoryfoster/skills#345 closed 2026-11-02, merge4a2ec874, and vendored here:.skills/deploy_commandsupport is in the vendoredshipping-work[measured 2026-11-06]; Package broker's release deploy (CannObserv/broker#22) into the FastAPI skills gregoryfoster/skills#469 closed 2026-10-28 as their duplicate) · shipping-work*: a.skills/deploy_commandknob — ship through the project's deploy script instead of "migrate, then systemctl restart" gregoryfoster/skills#345 lands only in the unifiedshipping-work(decided upstream 2026-10-31), so this also came after Adopt the unifiedshipping-workskill: symlinks and configuration, not overrides (gregoryfoster/skills#537) #174 (met 2026-11-02:shipping-worklinked) · after Host disk GC: adopt provisioner#61's canonical GC + headroom sensor on co-replicator (root hit 97%) #148 (releases need disk headroom) · unblocked 2026-11-14: deploying-releases: one vendored release-deploy core (fresh install and retrofit); init-project-fastapi delegates to it gregoryfoster/skills#372 Part A (thedeploying-releasescore) shipped upstream atf728abd5(announced on this issue, issuecomment of 02:20Z); not yet vendored:main's submodule is4f26e316, the plan commit, 15 behind [measured], so the bump is its first step · after Catch up co-core 0.45 → 0.46 (co-pypi index; closure diff version-only) #199 (operator, 2026-11-14) · coordinator note 2026-11-14 (issuecomment-6881112594): root 81%, 3.6G free,.venv309M per release, soKEEP=5≈ 1.5G → about 89%, past the 84.0% patching ceiling and the sensor'salert_at85; sizeKEEPin the design note [measured] · next (2026-11-14) · brief on the issue (issuecomment-6883663158), pre-flighted on91e87dc: body stale (names Package broker's release deploy (CannObserv/broker#22) into the FastAPI skills gregoryfoster/skills#469); brief = that comment + the 02:20Z comment +retrofit.md; step 0 bumps the vendored skills to skillsmain(≥cd7471ff; was55ede55f, 11 pastf728abd5: Part B and the co-pypi assets); the co-pypi wrapper stays asUV_WRAPPER(swap deferred to Adopt the shared co-pypi wrapper; move the wheel fetch to its own upstream-ready script #201, operator 2026-11-14);Settings.blob_dir's relative defaultblobs(config.py:152) would land inside a read-only release; the Refuse production write identity on a non-main checkout #52 guardcheckout.py:74changes in the same step; stop for the design note, then before the cutover (host session, operator's root writes) · restarts owed (the cutover) · verified 2026-11-14:91e87dc..b61703a(6 commits, fast-forward: bump9575b27to skills55ede55f, design docaad1fbbapproved with all six decisions as recommended, code1e8196f, sweep444c2cb,M=98f7bf2,C=b61703a); the implementer's pickup corrections, kept: a release is ~247M (--no-dev),KEEPcounts beyond live + previous, soKEEP=0(my 17:01Z arithmetic corrected in place); the wrapper needs its key exported by the hooks;BUILD_IDis now a sha12; the branch-switch proof replaced by a/procproof; the oldest deployable commit isM; cutover 21:2x–23:02Z with the operator's root writes, rollback drillC→M→C, each verified by the core [stated by the implementer]; CI green by job on98f7bf2andb61703a; coordinator:live→releases/b61703aee439, releases98f7bf25c47candb61703aee439root:root 555248M each; worker pid 59004 cwd the release, 0 mappings and 0 fds under/home;BUILD_ID=b61703aee439, NRestarts 0, handler silent; both installed units byte-identical to the live release; core marker version 1;KEEP=0;.skills/deploy_commandsetsid -w scripts/deploy.sh; replay 509/10, all threeenabled, 20 INFO only; default 2028 + 1 skip clean and loaded,tests/deploy32,tests/ci44, ruff clean [measured]; agent memory retired by the coordinator (design note §6 list); follow-ups Run replicator.service as a dedicated user, not exedev (#151 follow-up) #202, Drift check between deploys for the installed units and host configs (#151 follow-up) #203, deploying-releases: the service template starts Python through the live link, so a draining process imports the next release gregoryfoster/skills#649, deploying-releases: hardlink identical files between releases to cut per-release disk cost gregoryfoster/skills#650; closed by the coordinator. Root 83.72% after: Grow co-replicator's root disk (83.7% after the release cutover; patching gate at 84.0%) #204 - Adopt the shared co-pypi wrapper; move the wheel fetch to its own upstream-ready script #201 Adopt the shared co-pypi wrapper; move the wheel fetch to its own upstream-ready script · S–M · operational conformance · filed 2026-11-14 (operator): retrofit.md §1 replaces a repo's wrapper with
deploying-releases' copy (init-project-fastapi: replace the find-links wheelhouse with the co-pypi Artifact Registry index (Path C) gregoryfoster/skills#632,cd7471ff), which has nowheelsmode [measured]; the closure diff stays (6 of the last 8 pin moves foundchangedclosure files; Catch up co-core 0.40 → 0.41 (BlobStore reporting twins: keep SweepSafeLocalBlobStore's retry where calls reach it) #179's twin-verb finding changed what shipped) · built so the fetcher (and later the closure diff) can be proposed upstream as-is: noREPLICATOR_names, runs through the shared wrapper, self-contained test; no upstream filing without the operator's consent ·REPLICATOR_CO_PYPI_CREDENTIALS→CO_PYPI_KEY(default/etc/replicator/co-pypi-reader.json,root:exedev 640) · the issue body is the brief · after Run replicator's unit from a release, not/home/exedev/replicator(CannObserv/broker#22's rollout, replicator first) #151 · restart likely none [traced] · after Grow co-replicator's root disk (83.7% after the release cutover; patching gate at 84.0%) #204 too (its ship deploys a release); For Adopt the shared co-pypi wrapper; move the wheel fetch to its own upstream-ready script #201 note on the issue 2026-11-14 (issuecomment-6886156597):scripts/deploy_hooks.sh:41-48(theUV_WRAPPERcomment and the key export) is in scope; the ship's deploy is the end-to-end proof; vendored55ede55falready hascd7471ff· next (2026-11-15) · pre-flight on7706bf0(issuecomment-6886810710): upstream8f2dbfaeis 21 past the vendored55ede55f, wrapper still version 1 with the same sha256; recommend bumping first (brings Wave 8 set D); new literal sites:tests/deploy/test_deploy_hooks.py:66-71(asserts the key export; goes red, expected), the manifest's:171, the closure diff's help and error strings and their tests; the ship's deploy through the new wrapper is the end-to-end proof · verified 2026-11-15 (issuecomment-6889079646):7706bf0..a8900f5(the hook's bump4ae39d3, then PR #201 Adopt the shared co-pypi wrapper; fetch wheels with their own script #205's 10 commits, fast-forward); deployed byscripts/deploy.shat 03:52Z, the first production mint from the default key path with no*CO_PYPI*set; the operator removed/etc/replicator/.envlines 6–7; the fetcher proposed upstream as deploying-releases: ship a co-pypi wheel fetcher beside the wrapper, from CannObserv/replicator gregoryfoster/skills#652 and the closure diff's state posted on Shared import-closure diff for wheelhouse-pinned libraries (co-core pin moves), lifted from CannObserv/replicator gregoryfoster/skills#454 (operator's go-ahead) [stated by the implementer]; CI green by job ona8900f5; coordinator:live→releases/a8900f5d15a2(previous7706bf05cc16), pid 49915 cwd the release, 0 maps under/home, NRestarts 0, handler silent, 18 INFO, replay 511/10; the liveco-pypi.shequals main's, markers version 1 (ff3579e0…),install.sh --co-pypi --checkcurrent 1 1; env file has noCO_PYPI/wheelhouseline,root:exedev 640; the new fetcher underenv -ifetched 12 wheels, 0.46.1 ×3 matchuv.lock, the real-wheel closure test 38 passed, the 0.45.0 → 0.46.1 dry run (909 B) is in the pin comment line for line; default 2044 + 1 skip clean and loaded,tests/deploy+tests/ci76, ruff clean [measured]; closed by the implementer - Run replicator.service as a dedicated user, not exedev (#151 follow-up) #202 Run
replicator.serviceas a dedicated user, notexedev· M · follow-up of Run replicator's unit from a release, not/home/exedev/replicator(CannObserv/broker#22's rollout, replicator first) #151 (D1; deploying-releases D2 starts a retrofit onexedev) · filed 2026-11-14 by the implementer (operator's consent) · touches both units'User=,StateDirectory=ownership (incl.ms-playwright), the/run/replicatorchown,browser_cgroup.py's delegated subtree,/etc/replicator/*groups and the core'scheck_users· Decided 2026-11-15 (operator): the shared-group model: areplicatoruser and group,/etc/replicatorroot:replicator 750/640,exedevjoins the group soscripts/load_env.shkeeps working · next (2026-11-15): the worker runs asexedev, which has passwordless sudo anddocker, withNoNewPrivileges,ProtectHomeandProtectSystemall off [measured], so a compromised fetch or render is root and can read the repo.envPATs · brief and pre-flight ona8900f5(issuecomment-6889079947): the user must exist before the deploy (the core'scheck_usersrefuses otherwise);StateDirectory=chown,Delegate=to the new user, the failure handler'sUser=,rehearse_reconnect.sh:139, four docs; hardening (NoNewPrivileges,ProtectHome) in scope if-m browserand a live render prove the sandbox survives · verified 2026-11-15 (issuecomment-6897781795):a8900f5..f2a4333(13 commits, fast-forward; 8 CR fixes); the operator-authorized account (replicator997:988, home/nonexistent) andchgrp -R replicator /etc/replicatorat 05:07Z; deployed undersg replicatorat 05:24:47Z; a production capture at 05:28:57Z (operator-triggered, 1280x3385); a reboot at 05:33:12Z so sessions carry the group [stated by the implementer]; CI green by job onf2a4333; coordinator, after the reboot:live→releases/f2a43335c6b0(previousa8900f5d15a2), pid 443 user/group/supgroupsreplicator,NoNewPrivs: 1, cwd the release, 0 maps under/home, NRestarts 0;User=replicator,NoNewPrivileges,ProtectSystem=strict,ProtectHome,PrivateTmpon both units;/run/replicatorroot:root, onlybuild-id.env;worker readyf2a43335c6b0, all threeenabled, replay 507/10, INFO only, handler silent; 1stored a capturesince 05:24Z; asreplicatorsudo -nasks a password and/home/exedevis denied;/var/lib/replicatorallreplicator;/etc/replicator750 root:replicator, no other-bits; this session has the group andload_env.shloads withoutsg; default 2062 + 1 skip clean and loaded, host deploy tests 181, ruff clean [measured]; closed by the implementer - Drift check between deploys for the installed units and host configs (#151 follow-up) #203 Drift check between deploys for the installed units and host configs · S–M · follow-up of Run replicator's unit from a release, not
/home/exedev/replicator(CannObserv/broker#22's rollout, replicator first) #151 (D9) · filed 2026-11-14 by the implementer (operator's consent) · today only the host-runtest_the_installed_unit_is_the_live_releasesand the core'sHOST_CONFIGScompare at the next deploy catch drift; broker's pilot runs a drift timer · adopt the skill's once it ships, or port broker's; enabling is the operator's (ENABLE_NEW) · the known drift (99-co-replicator-memory.conf's comment) was discharged 2026-11-15 · parked 2026-11-16 (operator): adopt deploying-releases: ship the drift check once (live lags main in code that runs), as the core ships gregoryfoster/skills#641's shared drift check rather than port broker's; asked there (issuecomment-6906658975) that it also comparedeploy/units andHOST_CONFIGSwith the live release; park note issuecomment-6906659236 · re-entry in Parked - Adopt the unified
shipping-workskill: symlinks and configuration, not overrides (gregoryfoster/skills#537) #174 Adopt the unifiedshipping-workskill (Gate topology: run each check once per change (xdist, tree stamp, worker resolver) and fold shipping-work-* into one skill gregoryfoster/skills#537) · M · filed 2026-10-31 (upstream's adoption ask; release 2 deletes the variants once all 13 cohort repos move) · prerequisite of Run replicator's unit from a release, not/home/exedev/replicator(CannObserv/broker#22's rollout, replicator first) #151 (shipping-work*: a.skills/deploy_commandknob — ship through the project's deploy script instead of "migrate, then systemctl restart" gregoryfoster/skills#345 lands only in the unified skill) · ordered first (operator, 2026-11-01) · brief 2026-11-01 on the issue: the submodule isb386bd2b, upstream26d5f53eis 136 commits ahead and contains0aedfd4d; no overrides exist; 11 tracked lines name the variant, 4 of which hold the doctor; the unifiedpre-ship.shcarriesload_env, differing in one line (digit-led keys skipped; neither env file has one); step 6's CI census leans to waiting for upstream · no restart · verified 2026-11-02:2a6e0f2..337623a(the hook'sa3c8b10made the bump to26d5f53eand doctor2026-10-31-1; thena25701b,de4111d,be2a67c, fast-forward; then the hook's337623ato170af367); CI green by job onbe2a67cand337623a; 1819/0 on337623a; bothshipping-worklinks are vendor symlinks, no override,git grepfor the variant path exits 1, doctor--check-onlyexits 0;rcli PINGfrom anenv -ishell in/→ PONG; the operator's global instructions now callscripts/load_env.sh; no restart (build2a6e0f2); gate comparison on the issue; closed. Corrected in place: the brief's "breaks at release 2" (it broke at this merge, fixed at ship); the brief misseddocs/SKILLS.md:58 - chore: bump astral-sh/setup-uv from 10.3.2 to 10.3.3 #176 Dependabot:
astral-sh/setup-uv10.3.2 → 10.3.3 inci.yml· S · opened 2026-11-02 by Dependabot (Adopt Dependabot: nothing proposes crossing this repo's dependency caps #76's config) · SHA68559e9is tag v10.3.3 [measured]; the diff is the fouruses:lines; CI green by job onbf040c1· Decided 2026-11-03 (operator): the coordinator merges, as chore: bump astral-sh/setup-uv from 5.4.2 to 10.3.2 #168 was · merged 2026-11-03 as64c92c0(a merge commit on0a3c5af;tests/ci20 passed locally); CI green by job on64c92c0(test, lint, gcs, browser); GitHub marked the PR merged 15:37:12Z; no restart - Egress guard: add the IANA special-purpose ranges the default deny table lacks #161 Egress guard: add the IANA special-purpose ranges the default deny table lacks · S · filed 2026-10-27 (operator), from Egress guard: IPv6 forms embedding a private IPv4 address (NAT64, 6to4, ::/96) pass the deny table #156 · Decided 2026-10-30 (operator): all 12, each with its RFC; one doc clause per class list · verified 2026-10-31:
be3d735..37b6eca(the hook'sd37f775, then 5 commits, fast-forward, no PR); CI green by job on37b6eca(main's tip); live since 16:00:07Z, build37b6eca, NRestarts 0, 0 WARNING/ERROR, 0 live refusals; all 12 rows plus NAT64 and 6to4 embedded forms refused in the service venv (implementer 5, coordinator 10), public allowed; 1819/0; exact budgets match (outcome reference 9869, STREAMS 9800); host has nothing in198.18.0.0/15; left-out ranges listed on the issue; closed. Watcher's copy filed as Egress guard: add the 12 IANA special-purpose ranges Replicator's #161 added (documentation, benchmarking, ORCHIDv2, SRv6 and the rest) watcher#412 (operator) - Context budget: COMMANDS.md and ENVIRONMENT.md are over the 10k per-doc budget #162 Context budget: COMMANDS.md (10096) and ENVIRONMENT.md (10051) are over the per-doc budget · S · filed 2026-10-27 (operator), from Egress guard: IPv6 forms embedding a private IPv4 address (NAT64, 6to4, ::/96) pass the deny table #156 ·
curating-contextwith per-file telemetry · after After F1: state the fetch-policy producer as Archiver, retire the "until F1" wording #144 (both touch the SocratiCode manifest) ·context-cadence.ymlwarns only on AGENTS.md · next · brief on the issue 2026-10-30: exact counts unchanged (COMMANDS 10096, ENVIRONMENT 10051); DEPLOYMENT.md now 9664 after Adopt Dependabot: nothing proposes crossing this repo's dependency caps #76; corrected in place: a per-doc check can't go in the generatedcontext-cadence.yml· verified 2026-10-30:cda2861..ec96147(the hook's4cac2a2, then 5 commits, fast-forward); CI green by job onec96147(4cac2a2was red ontest, fixed by660f49b); exact: COMMANDS 9673, ENVIRONMENT 8931, no dead links; suite 1756/0; re-embed done by the session server's watcher; no restart - Load env files by parsing, not set -a; . — the documented load line sources the repo .env #170 Load env files by parsing, not
set -a; .· S · filed 2026-10-30 (upstream follow-up, init-project-fastapi: the templates source .env with set -a, and the generated unit loads the repo .env in production gregoryfoster/skills#517) · brief 2026-10-30 on the issue · four doc sites plus a test comment; both env files parse identically [measured] · hazard: AGENTS.md 5978 of 6000 · verified 2026-10-30:ec96147..ccd034f(7 commits, fast-forward); CI green by job onccd034f; suite 1783/0; AGENTS.md 5960;rcli PINGfrom anenv -ishell → PONG;scripts/load_env.shplus a repo-wide guard (tests/test_load_env.py); the operator's global instructions moved too; no restart - Drift test: hold every documented SocratiCode cap to the hook's default, and document the capped update #171 Drift test over every documented SocratiCode cap, plus the capped
updateline · S · filed 2026-10-30 (upstream follow-up, init-socraticode: MemoryMax= caps don't bound swap - add MemorySwapMax=0 on hosts with swap gregoryfoster/skills#340) · re-scoped 2026-10-30 (operator): its three steps landed via the hook's4cac2a2and Context budget: COMMANDS.md and ENVIRONMENT.md are over the 10k per-doc budget #162 (660f49b,f14733b); left isMANUAL_CAPreading only COMMANDS.md (#286 Resize to 8 GB: 4 G swap at swappiness 10, host wording updated archiver#292's shape) · after Load env files by parsing, not set -a; . — the documented load line sources the repo .env #170 · brief 2026-10-30 on the issue: exclude.skills/(old caps kept as warrants), identify a cap byMemoryMax=, notsystemd-run, share Load env files by parsing, not set -a; . — the documented load line sources the repo .env #170's_tracked_lines· verified 2026-10-30:ccd034f..be3d735(6 commits, fast-forward); CI green by job onbe3d735; suite 1793/0; COMMANDS 9800; 7 caps held to the hook's default,.skills/excluded,tests/_tracked.pyshared; re-embed done by the session watcher; no restart - Disk GC VERSION 4: update co-replicator from the vendored pin (skills#551) #164 Disk GC VERSION 4: update co-replicator from the vendored pin (retitled 2026-11-08) · S · filed 2026-10-28 (upstream's adoption ask) · hazard here: the VS Code client is on 1.140.0, so VERSION 3 keeps every newer staged build (641M per release; posted on maintaining-host-disk: editor-builds.md difference 1 is measured — a removed staged VS Code build came back in 5.5 h (co-replicator) gregoryfoster/skills#504) · Decided 2026-10-28 (operator): the operator updates the VS Code client first, then this runs after Adopt Dependabot: nothing proposes crossing this repo's dependency caps #76 as ordered; verify the live build is current before the brief · pre-flight on the issue 2026-10-28 · step 1 already done by the hook:
main's submodule5c3e486econtainsb1d6e303· the client was still on 1.140.0 at 2026-10-30T00:2xZ (2c33b50re-staged) · Decided 2026-11-06 (operator): rescoped toVERSION4, which keeps only the newest idle build newer than the live one (the image repo, released 2026-11-01). Waits on maintaining-host-disk: pin the disk GC's VERSION 4, and update editor-builds.md row 1 gregoryfoster/skills#551 (the skill's pin move, built and held on a local branch; upstreammainand vendored227de51estill pin 3) reaching upstreammainand being vendored · the client moved: livea717861(1.145.0), with 1.145.1 and 1.145.2 staged [measured 2026-11-06T18:4xZ]; re-read on the issue (issuecomment-6692706156) · upstream nudged 2026-11-06 (operator): maintaining-host-disk: pin the disk GC's VERSION 4, and update editor-builds.md row 1 gregoryfoster/skills#551 (issuecomment-6697525603) · unblocked 2026-11-08: maintaining-host-disk: pin the disk GC's VERSION 4, and update editor-builds.md row 1 gregoryfoster/skills#551 shipped upstream 2026-11-07 (b51f3c24), vendored by the hook inedb63bc;_pin.shreadsDISK_GC_VERSION=4; the host has 2 [measured] · pre-flight pull and plan read-only:update2 → 4, nothing removed,journald_restart: false, units and drop-in byte-identical, config keys unchanged · ordered (operator, 2026-11-08) · brief on the issue (issuecomment-6728461637): the sudo apply on the owner's approval in its own thread; gate on the install and the first run's exit 0; the "newest kept" line waits for a staged build (Watching) · noreplicatorrestart · verified 2026-11-09: applied 00:27Z on the owner's approval in its thread, no repo change; coordinator read-back:VERSION4, MANIFESTsha256 --strictclean and matching the pin (89059897…), a fresh pull's re-plancurrent4/4, both timersenabled, the hand-started runResult=successandstate.jsonexit0, the sensorchecked in ok (ok): HTTP 202; root 74% before and after;build-id74c69a7, NRestarts 0; the "newest kept" line is under Watching; closed by the implementer (record on the issue) - Fetch policy: retire the unordered path once nothing can publish a frame without a generation #163 Fetch policy: retire the unordered path · S (trigger A) · un-parked 2026-11-06: trigger A met, Revoke Watcher's content.fetch-policy publish grant broker#95 closed 2026-11-03 (Watcher's grant revoked) [measured]; a docs-only trim of about 25 tokens (
POLITENESS.md:176-178,policy.py:364, the module docstring) · trigger B (the path's removal) is still parked: its first condition, CannObserv/cannobserv#579's requiredgeneration, arrives with Catch up co-core 0.40 → 0.41 (BlobStore reporting twins: keep SweepSafeLocalBlobStore's retry where calls reach it) #179; retention andFetchPolicyState'sNoneremain · brief 2026-11-06 on the issue (issuecomment-6692706556): two sites, not three (the module docstring names Watcher's pre-F1 frames, cause 1, which is B's); Catch up co-core 0.40 → 0.41 (BlobStore reporting twins: keep SweepSafeLocalBlobStore's retry where calls reach it) #179's For Fetch policy: retire the unordered path once nothing can publish a frame without a generation #163 notes carried there · no restart · trigger A verified 2026-11-06:beb6d3b..7460a5c(2 commits, comments and prose only, fast-forward); CI green by job on7460a5c; coordinator re-runs: default 1907/0,tests/test_boundaries.py97, ruff clean; POLITENESS 8128 → 8106 exact; no restart (build-id74c69a7) · the issue stays open for B, parked below (issuecomment-6697158717) - CI: bound every job and network install step with timeout-minutes (an apt hang held test for ~2.5 h) #180 CI: bound every job and the network install steps with
timeout-minutes· S · filed 2026-11-06 by Catch up co-core 0.40 → 0.41 (BlobStore reporting twins: keep SweepSafeLocalBlobStore's retry where calls reach it) #179's implementer:test's "Install redis-server" hung ~2.5 h on the runner;ci.ymlsets no timeout (GitHub's 360 min default) · measured job maxima on the issue (lint 3.7, test 4.5, browser 14.3, gcs 3.9 min) · brief 2026-11-06 on the issue (issuecomment-6697159106): a red-firsttests/ciguard, measured bounds with their measurements beside them,context-cadence.yml:35the precedent · no restart · verified 2026-11-07:7460a5c..7b21e86(12 commits, fast-forward); jobs lint 15, test 30, browser 30, gcs 15, both apt steps 15, the arithmetic inci.yml's comment;tests/ci/test_workflow_timeouts.pyred-first; CI green by job on7b21e86; coordinator re-runs:tests/ci27, ruff clean, default 1914 (one earlier run hit the watched flakecorrected 2026-11-08: that run's 5 failures wereload_env.shleaking production flags into pytest, Tests: conftest's PRODUCTION_ENV doesn't scrub REPLICATOR_CAPTURE_ENABLED or REPLICATOR_BROWSER_FETCH_ENABLED, so pytest after load_env fails 5 tests #182; 14 clean-shell runs); no restart; closed by the implementer (issuecomment-6713209502) - Tests: fold tests/test_ci.py into tests/ci/ (CR 7 of #180) #181 Tests: fold
tests/test_ci.pyintotests/ci/· S · filed 2026-11-07 (operator), CR 7 of CI: bound every job and network install step with timeout-minutes (an apt hang held test for ~2.5 h) #180 · 10 tests move, no assertion changes; doc sitesdocs/TESTING.md:251,docs/ARCHITECTURE.md:73· the issue body is the brief · no restart · verified 2026-11-08:7b21e86..83af705(the hook'sedb63bc, then 5 commits, fast-forward);tests/ci/test_workflow_runs.py, 10 tests, its own module; brief review corrections accepted (scope CI: a main commit can land with no CI signal — the concurrency group holds only one pending run #44/Run the broker-OOM suite in CI — it is the one integration file that needs no VM broker #83/Viewport capture: consume the capture command, render headless, store, publish the stored-capture fact #138, two more docstring sites); CI green by job on83af705; coordinator re-runs:tests/ci37, ruff clean, default 1913 + 1 environmental skip in a clean shell; no restart; closed by the implementer (issuecomment-6724331474) - Resolve co-core from the co-pypi Artifact Registry index, retiring the GCS wheelhouse (#169 part 1) #183 Resolve co-core from the co-pypi Artifact Registry index, retiring the GCS wheelhouse · M · split from Dependabot: add the uv ecosystem once the private wheelhouse has an index it can resolve #169 2026-11-08 (operator) · Watcher's precedent Dependabot for the uv ecosystem: the wheelhouse blocks the updater watcher#358 part 1;
co-pypi-reader@reads the index [measured] · design doc first; isolated venv; the closure diff's wheel source must be replaced ·deploy/replicator.servicechanges, restart owed · brief on the issue 2026-11-09 (issuecomment-6730110479): keeps 0.41.0; the host's reader key isREPLICATOR_WHEELHOUSE_CREDENTIALS, not ADC (that's the writer); start step A (no sync) or B (--locked --offline, never--check); 9 surface sites the body missed; design doc first · pre-flight refreshed 2026-11-10 onc74be97(issuecomment-6773881481): nothing in86c8986..c74be97touches its surface;browser's two known flakes are fixed · verified 2026-11-10:c74be97..294e489(14 commits merged asbc4c5ec, then294e489; designdocs/plans/2026-11-10-co-pypi-index-design.md, approved items 1–5);co-pypi[[tool.uv.index]]explicit,scripts/co-pypi.shmints fromREPLICATOR_CO_PYPI_CREDENTIALSonly (renamed in/etc/replicator/.env, operator-authorized) and starts uv from an allowlisted environment; start step A (no wheelhouseExecStartPre); closure diff takes--wheelsfrom the index;docs/DEPENDENCIES.mdnew; sha256 of all six 0.41.0 files equal in lock, index and.wheelhouse/[measured by the implementer]; CI green by job onbc4c5ecand294e489(run 42877329370); coordinator: live since 18:36:53Z, buildbc4c5ec, NRestarts 0, installed unit =deploy/,uv pip show0.41.0 ×3, replay 505/10 hosts, 12 blobs stored since, 0 WARNING/ERROR;.wheelhouse/gone; onlyREPLICATOR_CO_PYPI_CREDENTIALSin the env file (name only read); default 1959 + 2 skips in a clean shell and in a loaded one withREPLICATOR_REDIS_URLunset,tests/ci41, ruff clean; closed. Follow-up: test_inflight's env-file test fails in a loaded shell and prints the broker URL with its password #187. Limits under Accepted limitations - docs: retire co-gcs-blob-reader as the co-gcs-blobs reader (watcher#420) #185 docs: retire
co-gcs-blob-readeras theco-gcs-blobsreader · XS · docs · filed 2026-11-09 from a Watcher session (cross-repo policy) ·docs/INFRASTRUCTURE.md:134(Reader) and:136(Consumer key, already false: Watcher's key was shredded 2026-11-09) [measured] · unblocked 2026-11-15: Delete the co-gcs-blob-reader service account and its keys (after 2026-11-16 soak) watcher#420 closed 19:56Z (date gate waived by the operator; bindings removed, account deleted, read-back 0/0/0 [measured by the operator]; undo until ~2026-12-15); docs: retire Watcher's co-gcs-blob-reader grant on co-gcs-processor (watcher#420) processor#130 closed · pre-flight issuecomment-6903568197 addsINFRASTRUCTURE.md:191("Watcher's SA … opens the bytes", false since Jump: delete the notification subsystem, the notifier client and their configuration watcher#379) andsrc/worker/main.py:260's docstring; the Reader row needs the operator's read-onlyget-iam-policy(no identity here can list it), else mark it retired · INFRASTRUCTURE.md at 9999 of 10000, must not grow · no restart unless the docstring changes · verified 2026-11-16 (issuecomment-6928781031):59fb240..d549a06(the hook's6ee9b95+ 3 commits, fast-forward); CI green by job ond549a06and6ee9b95; the Reader row became Readers from the operator's liveget-iam-policy(co-gcs-processor-writer,co-gcs-publication-writer,objectViewer), the Consumer key row deleted,:191's Watcher sentence dropped,main.py's docstring now says "a consumer"; Done-whengit grepclean; INFRASTRUCTURE.md exact 10101 → 10011 (estimate 9925); coordinator 2062/1 clean and loaded, ruff clean; both indexes current; not deployed (operator): live59fb24082f7etrailsmainby text only [measured]; closed by the implementer - Switch to shipping-work's pr merge mode (.skills/merge_mode) #208 Switch to shipping-work's
prmerge mode (.skills/merge_mode) · S · filed 2026-11-17 (operator), from the idle check: the refresh hook pushed85f8d7bstraight tomain[measured]; no knob, no rulesets · three writers move: shipping-work, the refresh hook (one PR fromskills/submodule-refresh),context-cadence.yml(re-render;--checkgoes STALE otherwise) · pre-flight issuecomment-6952757211: the Actions "create and approve pull requests" setting is unreadable by the PAT, so the operator checks it;.skills/deploy_commandset, so production issues takeRefs; recommends merge commits;docs/SKILLS.md:258-280anddocs/DEPLOYMENT.md:268to update · an enforcing ruleset is out of scope (operator's later decision) · no worker code, deploy optional · verified 2026-11-17 (issuecomment-6955855095):85f8d7b..a072ba5(7 commits, PR #208 Switch to shipping-work's pr merge mode #209 merged on GitHub with a merge commit, the first, operator's choice); CI green by job onaa1bb8aand ona072ba5as apushrun, so a GitHub merge gets the rundeploy.sh's gate needs [measured];merge-mode.sh→pr,install-cadence.sh --checkexit 0 (merge mode: pr,re-render: none), doctor clean; implementer's pre-flight corrections accepted (the installer reads the primary's knob, so theprrender was made in a scratch clone;DEPLOYMENT.md:57andSKILLS.md:248-262added; rollback targets are merge commits); coordinator 2062/1 clean and loaded,tests/ci44, ruff and lock clean; not deployed (operator): live59fb24082f7e; Actions may open PRs (operator, issuecomment of 16:38Z) [stated]; filed context-cadence.yml still pushes to main under merge_mode pr: re-render it archiver#525, install-cadence.sh: let the branch that flips .skills/merge_mode carry its own re-render (and doc-check recognize it) gregoryfoster/skills#674; CR 6 (SHA-pinactions/checkout@v7incontext-cadence.yml) held for the operator; closed by the implementer - Catch up co-core 0.48 → 0.50 (0.50.0) #211 Catch up co-core 0.48 → 0.50 (0.50.0) · S · filed 2026-11-18 (operator), from the 05:47Z idle check: v0.49.0 and v0.50.0 tagged 2026-11-17T17:32Z and 21:44Z, both on the co-pypi index ×3 [measured, unbounded]; pin
>=0.48.0,<0.49· closure diff 0.48.2 → 0.50.0 from index wheels: 58/58, 6 not identical,changeddrivers/gdrive.py(Drive, refused here) anddrivers/transport.py(0.49's deadline-race fix; noresilient_adaptercall site here,GcsBlobStoreusesgoogle.cloud.storagedirectly) [measured, traced];changes.pydocstrings-only,generationstill optional (Fetch policy: retire the unordered path once nothing can publish a frame without a generation #163 B stays parked) · the issue body is the brief, with every literal site · ships through a PR (Refs), the first deploy of a GitHub merge commit · isolated venv · restart owed · verified 2026-11-18 (issuecomment-6982039306):796186f..753ecb9(b464c83+ PR #211 Catch up co-core 0.48 → 0.50 (0.50.0) #213's merge commit); CI green by job on both,753ecb9's ownpushrun 44323517457; first deploy of a GitHub merge commit: live753ecb9609a0since 14:25:36Z,replicator, NRestarts 0, 0.50.0 ×3 in the release venv, replay 507/10 hosts, all INFO [measured]; coordinator 2062/1 clean and loaded, deploy tests 181, ruff and lock clean; manifest re-embedded; brief corrected in place twice (InformationModel.sourceis co-core's own;transport.pyloads at import); closed by the implementer
Waits
| Ref | Who waits | On what | Asked | Unblock condition |
|---|---|---|---|---|
| gregoryfoster/skills#372, gregoryfoster/skills#345 | #151 | the release-deploy core and the deploy_command knob |
on gregoryfoster/skills#469, 2026-10-23 | met 2026-11-14: Part A shipped (f728abd5) and vendored at 55ede55f (9575b27); #151 shipped on it. gregoryfoster/skills#372 stays open for its own close-out |
| CannObserv/cannobserv#554 | #169 (the uv half, parked) | an index Dependabot's uv updater can resolve | on the issue | the index is served and declared as a registries: entry |
| CannObserv/broker#116 | us (verification) | Archiver's half: broker reads the group again after the first capture command. Replicator's half read back clean on 2026-10-27 | 2026-10-26, on the issue | broker's second read-back, after CannObserv/archiver#380's first command |
| CannObserv/archiver#380 | us (verification); them for the rest | Archiver issuing content.capture |
go-ahead posted 2026-10-26 | the first stored a capture and published capture_stored in our journal |
| CannObserv/archiver#411 | them | #138 (met) and their own design | — | — |
| CannObserv/archiver#379 | them | #63 | — | met 2026-11-06: deployed and announced (issuecomment-6677665321) |
| CannObserv/watcher#414 | them | #63's support announcement (co-core's MUST: no issuer sets render before it) |
filed 2026-10-31 by the cannobserv coordinator | met 2026-11-06: announced (issuecomment-6677665074); the MUST is discharged |
| CannObserv/watcher#412 | them | #161 (met); hardening, their call | filed 2026-10-31 (operator), after CannObserv/watcher#402 (the #156 port, closed 2026-10-28) | — |
| CannObserv/cannobserv#602 | #63 | co-core's render field on content.fetch |
filed 2026-10-28 (operator) | met 2026-11-01: v0.40.1 released (tag 0106cf45) and on the index [measured] |
| gregoryfoster/skills#536 | a future adoption issue here | the reviewing-code* variants folded into one stack-detecting reviewing-code (gregoryfoster/skills#537's D6) |
raised by #174's implementer, 2026-11-01 | merged upstream and vendored: then file the adoption issue (operator), covering the skills/ pair and docs/SKILLS.md:58 |
| CannObserv/broker#85 (the Valkey 7.2 window) | #178 | broker grants `+client | setinfotoreplicator` in the window, then posts its notice |
answered 2026-11-03 on #178 |
| CannObserv/cannobserv#635 | them | ContentFetchCommandEmit refusing any header but user-agent beside render="browser" (#63's call 1A) |
filed 2026-11-06 (operator) | — (meanwhile Replicator refuses it as invalid_request_options) |
| gregoryfoster/skills#551 | #164 | the skill's pin move to disk GC VERSION 4 (only the newest staged VS Code build kept); built and held on wave8/551-version-4 since 2026-11-03, its held question answered "no re-stage" the same day |
nudged 2026-11-06 (operator, issuecomment-6697525603) | met 2026-11-08: shipped upstream 2026-11-07 (b51f3c24), vendored in edb63bc, _pin.sh reads DISK_GC_VERSION=4 [measured] |
| CannObserv/cannobserv#640 | #163 B (its argument) | co-core's FetchPolicyState docstring stops naming a still-publishing Watcher (stale since CannObserv/broker#95) |
filed 2026-11-06 (operator) | met 2026-11-10: v0.42.0 (ce0277f5) rewords it to name both sources of an unordered frame; in our floor since #184 (0.45.0 now) [read 2026-11-13]. B's other conditions are unchanged (Parked) |
| CannObserv/watcher#420 | #185 (the Reader row) | co-gcs-blob-reader@ loses its objectViewer binding on co-gcs-blobs and is deleted |
filed by Watcher 2026-11-09 | met 2026-11-15: closed 19:56Z; both bindings removed, the account deleted, read-back on the issue; undo until ~2026-12-15 (uniqueId on the issue) |
Parked
Each row is brief-grade: a trigger may reach an implementer with no hand-off. When any upstream release lands (a co-core minor, a skills release), re-read every row, not only the one whose trigger fired. Rows re-read 2026-11-16 [measured unless marked].
| Ref | Trigger (observable) | When it fires | Hazards | Evidence | Last re-read |
|---|---|---|---|---|---|
| #203 Drift check for installed units and host configs | gregoryfoster/skills#641 ships its drift check, or splits the host half into its own issue | becomes the adoption: install the module with its marker (adopt unmodified), set the per-repo inputs (RUNTIME_PATHS, the monitor id, the grace) in scripts/deploy_hooks.sh or its sibling, the units from the template; the operator creates the co-status monitor and check-in key, the tailnet rule to tag:status:9000, then ENABLE_NEW |
GitHub's unauthenticated 60/h per address is shared with the deploy's CI gate; a host-half verdict must page only for units (production not running what was deployed), not for a host config waiting on a hand cp |
the scope ask issuecomment-6906658975 (2026-11-16), no reply yet; today the core installs differing units at each deploy and notes differing HOST_CONFIGS |
2026-11-16 |
#169 Dependabot: add the uv ecosystem |
all three: CannObserv/cannobserv#615 decides the PyPI name registration (or the operator accepts the PyPI-first limit); an org owner admits this repository to the private registry (cannobserv's docs/DEPENDABOT.md); #183's index move (met, bc4c5ec) |
add the uv ecosystem to .github/dependabot.yml with a registries: entry for co-pypi, keep the co-core current-minor caps as ignores, extend the tests/ci guards |
dependency confusion is the reason CannObserv/cannobserv#615 exists: unregistered names on PyPI; the updater must hold only the reader identity, never the writer's; a proposal crossing the pin's minor breaks the "pin the current minor" rule | CannObserv/cannobserv#615 open, its owner runbook paused 2026-11-14 (nothing uploaded); dependabot.yml's reason names this row |
2026-11-16 |
| #163 B Fetch policy: retire the unordered path | all three: co-core decodes FetchPolicyState.generation as required (still Generation | None at 0.48.2, changes.py:1974), or the operator chooses Replicator's own refusal; Archiver at co-core ≥ 0.41 (still >=0.38,<0.39); retention clear of Watcher's frames (only broker's XRANGE can say) |
remove the occurred_at branch, the ratchet, _log_out_of_order's None branch, POLITENESS :72, and the Watcher scenarios and docstrings (test_policy.py:540-542, test_policy_reader.py:578-657) |
a frame without a generation after the change must be refused loudly, never applied; the replay at boot reads the whole stream, so retention matters | _log_out_of_order's None branch: 0 firings since 2026-10-22; trigger A shipped 2026-11-06 (7460a5c) |
2026-11-16 |
| #64 WACZ/WARC capture research | CannObserv/archiver#408's design settles the producer and the trigger | rescope #64 to the design's Replicator half, or close it for the issue the design files; run the charter's three tests (docs/contracts/replicator-boundaries.md) on the capability first |
a new command stream means a design doc, a contract doc, broker grants (broker's) and a stream-kind entry; a WACZ is a browser workload, so it inherits #138's renderer confinements and memory bounds | CannObserv/archiver#408 open, 0 comments, last update 2026-10-20 | 2026-11-16 |
Sibling replies owed
- Per-source fetch timeout, carried to content.fetch's timeout_seconds archiver#377: answered 2026-10-24. The ceiling's doc, refused not clamped, and a flag: the journal shows Watcher's 1800 s doesn't reach
timeout_secondsas the issue states. - CannObserv/cannobserv#495: answered 2026-10-24. Replicator's precondition has been met since 2026-09-27 (Converge storage on the cluster content-addressed model #114); the remaining gate is Archiver's.
- [Epic] Headless Watcher and cohort boundaries watcher#368: notified 2026-10-28 that [Epic] content.fetch-policy producer moves to Archiver #135 is closed and Fetch policy: retire the unordered path once nothing can publish a frame without a generation #163 carries what remains.
- init-socraticode: mcp-driver.mjs has no update subcommand, so a capped codebase_update outside a session needs a shim gregoryfoster/skills#483: posted 2026-10-28 (operator): the capped
index .doesn't refresh a stale context artifact (mcp-driver.mjs:3011counts missing ones only); line reference corrected in place. - Pass info.registry's render mode through to content.fetch (cannobserv v0.40.1, cannobserv#602) watcher#414: announced 2026-11-06 (operator):
renderis honoured since Headless-browser fetch path for JS-rendered pages (re-sited from watcher#3) #63; a browser command may carry onlyuser-agent. - Per-source render mode (headless-browser fetch) as Archiver policy, carried to content.fetch archiver#379: announced 2026-11-06 (operator): policy notes for
browser(HTML only,load, one render at a time, the fingerprint discontinuity). - maintaining-host-disk: pin the disk GC's VERSION 4, and update editor-builds.md row 1 gregoryfoster/skills#551: nudged 2026-11-06 (operator): co-replicator waits on the
VERSION4 pin for Disk GC VERSION 4: update co-replicator from the vendored pin (skills#551) #164. Shipped 2026-11-07; nothing owed. - CannObserv/cannobserv#640: filed 2026-11-06 (operator): the stale
FetchPolicyStatedocstring sentence. - Viewport screenshots: per-source capture policy, issued to Replicator on a dedicated command stream archiver#380: posted 2026-11-12 from Store viewport captures in their own bucket, co-gcs-captures (archiver#380 D4) #189's thread (issuecomment-6813574780): prefixes, go-live 02:41:29Z, the first post-split capture. Operator approval given in that thread [stated]. Archiver's reader key and flip are theirs.
- Adopt patching-hosts: commit the knob, then take the first monthly run under the skill (gregoryfoster/skills#313) archiver#375: answered 2026-11-13 (operator consent, issuecomment-6865787520): Replicator isn't an archiver-API caller, so the patching knob's
caller CannObserv/replicatorand three README lines are stale. From Archiver patch run tonight 2026-11-13 (~22:00-23:59Z): ~1-2 min API outage at the reboot (archiver#375) #197, archiver's patch-window notice, closed as no impact.
Verification debt
| What is unverified | Why it was skipped | How to discharge it | Discharged by |
|---|---|---|---|
A real capture end to end (capture_stored on content.artifacts) |
Replicator can't XADD to content.capture; Archiver hasn't issued yet |
the journal shows stored a capture and published capture_stored, and XPENDING content.capture replicator.capture returns to 0 |
discharged 2026-11-11: 01M75H603VX307CE6YRMTJV3NR, stored a capture and published capture_stored at 14:46:26Z, Archiver applied it 6.4 ms earlier (their record on #138, now closed); a second at 19:11:09Z; the worker's own pending count across its groups reads 0 (scripts/inflight.py, 23:2xZ) [measured] |
replicator.capture read back, and ACL LOG clean |
XINFO GROUPS is denied to replicator |
broker posts its acladmin read-back |
discharged 2026-10-27: CannObserv/broker#116 (issuecomment-6450135332): the group exists at 0-0, created before any entry, 4 replicator readers blocked, no replicator entry in ACL LOG. consumers 0 is expected until the first delivery |
| The manifest's re-embed after #151 | another session's SocratiCode server held the index and context locks | the capped update from docs/COMMANDS.md, run from the main checkout, then codebase_context_search shows the new descriptions |
discharged 2026-11-15: the #201 implementer's server's watcher ran the incremental update; codebase_context_search returns #201's new DEPENDENCIES.md and COMMANDS.md text [measured by the coordinator] |
99-co-replicator-memory.conf's installed copy predates #112's comment edit; every deploy reports it |
values identical, so not urgent; a root write | the operator's sudo cp /srv/replicator/live/deploy/99-co-replicator-memory.conf /etc/sysctl.d/, then the next deploy reports no HOST_CONFIGS drift |
discharged 2026-11-15: installed by the #204 implementer with the operator's authorization at 00:30:30Z; its sha256 equals the repo's (6b42b277…), and the live values match [measured] |
The first pr-mode context-cadence run (#208) |
merged after the render; the schedule hadn't fired yet | a PR from context/weekly-measurement titled chore: weekly context measurement, no direct bot commit on main; merged with a merge commit (close and reopen for CI first). Expected 2026-11-17 ~21:00–22:30Z, else 2026-11-24 |
discharged 2026-11-18: the 2026-11-17T23:53Z run opened PR #210 (only the three .skills/context-* files), main did not move; its own pull_request run (actor github-actions[bot]) failed with 0 jobs; the coordinator closed and reopened it (operator's request), CI green by job on dba8b20, merged with a merge commit 796186f at 05:53:41Z [measured] |
The refresh hook's first pr-mode run (#208) |
today's ran at 08:43Z, direct | .git/skills-update.log shows the pr path at the first session on 2026-11-18; the first real PR from skills/submodule-refresh comes with the next upstream change; main gets no hook commit |
discharged 2026-11-18: the 13:12Z run took the pr path (log: restored both submodules to the recorded pointer, proposed d4b291c on origin/main) and opened PR #212, only the gregoryfoster-skills pointer, CI green by job; main got no hook commit [measured] |
Accepted limitations
- CR 10 of Viewport capture: consume the capture command, render headless, store, publish the stored-capture fact #138: each render's SOCKS listener is unauthenticated, and a local process could reach it. Accepted 2026-10-26.
- CR 11 of Viewport capture: consume the capture command, render headless, store, publish the stored-capture fact #138: Playwright's Node driver runs unconfined in the
worker/cgroup. A page can't reach it. Accepted 2026-10-26. - From Retention sweep can rmdir a shard the store just created, failing the write (transient; flakes test_run_dispatches_to_the_byte_path) #159, local backend only (production uses the object store), accepted 2026-10-26: a forward clock step larger than the grace can lose the one retry, and is then retried at the reclaim; emptied shards linger for one grace per level; the INFO line "a sweep removed the shard mid-store" is the only signal of a lost race.
- From Retention sweep can rmdir a shard the store just created, failing the write (transient; flakes test_run_dispatches_to_the_byte_path) #159:
REPLICATOR_BLOB_TEMP_GRACE_SECONDSis unbounded, and at0the shard grace and the one-retry argument both fail. Not filed: it needs a misconfiguration on a local-backend host. Re-entry trigger: a host moving to the local backend, or a boot-time bound proposed for another reason. Proposed 2026-10-26. - The co-core upstreaming of Retention sweep can rmdir a shard the store just created, failing the write (transient; flakes test_run_dispatches_to_the_byte_path) #159's retry is a draft on the issue, not filed. Worth it only if another cohort repo sweeps a co-core local tree. Operator's call.
- From Adopt Dependabot: nothing proposes crossing this repo's dependency caps #76: the
browserjob has noAssert wheelhouse auth is configuredstep, so an auth failure on a Dependabot run in that job can't be read against the verdict table. The other three jobs carry it. Re-entry trigger: the nextci.ymledit. Recorded 2026-10-30. Settled 2026-11-10 by Resolve co-core from the co-pypi Artifact Registry index, retiring the GCS wheelhouse (#169 part 1) #183:browsergained the WIF-provider assert. - From Adopt Dependabot: nothing proposes crossing this repo's dependency caps #76:
context-cadence.ymlruns onubuntu-latestwith a tag-pinnedactions/checkout@v5. It's generated byinstall-cadence.shand excluded from Dependabot, so a fix is an upstream installer change. Not filed: the operator declined, 2026-10-30. - From CannObserv/replicator#162: the skills-refresh hook commits and pushes a submodule bump without running the suite, and
4cac2a2pushed a redmain. That's an upstream ask (gregoryfoster/skills: run the repo's tests before pushing). Not filed: the operator declined, 2026-10-30. - From CannObserv/replicator#162:
install-cadence.sh's rendered workflow warns only on the policy file. The drafted upstream ask: warn on anydocs[].over_budget, and listnear_budget. Not filed: the operator declined, 2026-10-30. - From CannObserv/replicator#171 (CR 4): the drift test checks that documented caps agree, not that every documented SocratiCode launch is capped. A bare
node "$SOCRATICODE_DRIVER" index .line passes. The operator declined a guard, 2026-10-30. - From CannObserv/replicator#171: two tracked-file readers,
tests/test_doc_sensitive_paths.py::read_tracked_files(paths) andtests/_tracked.py(text). Merge them if a third caller appears. Recorded 2026-10-30. - From CannObserv/replicator#161, accepted 2026-10-31:
- a set
REPLICATOR_BLOCKED_DESTINATIONSreplaces the table and drops all 12 rows (the unwrap still holds); - a few globally reachable sub-entries are refused: PCP and TURN anycast in
192.0.0.0/24, the 6a44 relay192.88.99.2, and all of ORCHIDv2; 2001::/23is no row;- a site-chosen NAT64 prefix is still unchecked, as after CannObserv/replicator#156;
- the rest of the IANA registry (AS112, AMT, anycast services, DETs,
100:0:0:1::/64, the old ORCHID) is left out, with reasons on the issue.
- a set
- From CannObserv/replicator#172, accepted 2026-11-01: the pin comment's 0.39 entry is condensed under "Previously:", and the verbatim 0.37.0 → 0.39.1 closure report is in git at
37b6eca:pyproject.toml. - From CannObserv/replicator#174, accepted 2026-11-01:
.skills/import-targetsis unset, so the gate's import check skips:import srcloads onlysrc/__init__.py, and the suite imports every module. Recorded indocs/SKILLS.md.- No
.skills/test-workers: on 2 vCPUs the resolver goes serial on its own, and pytest-xdist isn't installed. - Step 6, the CI census, is deferred to upstream's prescriptions, because
ci.ymloverlaps CannObserv/replicator#151 and CannObserv/replicator#169. - Skipped: a committed red (the doctor holds while the old path is named; shown in-session), and the run-everything gate policy (nothing failed).
- From CannObserv/replicator#173, accepted 2026-11-02:
- a peer whose TLS breaks on every attempt now retries indefinitely (transient is exempt from the ceiling), as
ConnectErrordoes; before, it closedhandler_errorafter about 4 minutes. The outcome reference's "network error … retry indefinitely" row already covers it; - only
_fetchis covered: capture's TLS errors go through Playwright; - no real-TLS test in the suite (the guard refuses loopback); the escape was measured once by a scratch script.
- a peer whose TLS breaks on every attempt now retries indefinitely (transient is exempt from the ceiling), as
- From CannObserv/replicator#175, accepted 2026-11-03: production's refused-dead-letter alternation holds with about 59 s of margin at the defaults, but nothing refreshes a claim mid-handler, so a handler that takes about 59 s or more to reach a permanent failure under a sustained cap, or a
REPLICATOR_CLAIM_MIN_IDLE_MSset near the backoff base, reaches the restart [inferred from code, not measured]. Stated indocs/CONVENTIONS.md. The OOM test now takes about 5 s. - From CannObserv/replicator#177, accepted 2026-11-03: until CannObserv/replicator#63, a
render="browser"command is decoded and still fetched over HTTP. co-core's MUST (no issuer setsrenderbefore Replicator's announcement) guards it. - From CannObserv/replicator#63, accepted 2026-11-05 (design approved; listed on the issue):
- inherited from CannObserv/replicator#138: CR 10 (the per-render listener), CR 11 (the unconfined Node driver, which now relays the document, bounded by the in-page measure and
_raise_for_size), the CannObserv/replicator#156 and #161 guard limits, and unpaced subresources; - the document is the DOM at
load; later script content is absent until co-core has a wait field; the serialized DOM keeps the origin's<meta charset>; - unclassified (
handler_errorat the ceiling): an unbidden 304 (ERR_ABORTED), a navigation during serialization, an unmetload, and the deadline underbrowser(call 4); - a browser fetch can wait behind a capture inside its own 120 s deadline (call 5);
- a parked
browsercommand on a host with the flag off logs one WARNING per reclaim (CR 6, kept as the only signal); - skipped: no live
browsercommand yet, and no single end-to-end dev-worker run; the lock test proves itself by failing without the lock.
- inherited from CannObserv/replicator#138: CR 10 (the per-render listener), CR 11 (the unconfined Node driver, which now relays the document, bounded by the in-page measure and
- From CannObserv/replicator#179, accepted 2026-11-06:
- option (b) departs from co-core 0.41.0's
BlobStoredocstring (a wrapping subclass wraps the twins); before the first twin call, move the retry onto both twins and drop the plain-verb overrides; the guard goes red on a twin reference or override; - the guard sees literal names only (attribute or exact string) and scans
src/, notscripts/; - pre-existing, unfiled:
build_system.requires = ["uv-build>=0.12.0,<0.13"]warns under the host's uv 0.13.5.
- option (b) departs from co-core 0.41.0's
- From CannObserv/replicator#180, accepted 2026-11-07: no hang reproduced (the guard plus a green run is the evidence); the wheelhouse sync and
uv syncare bounded only by their job (30 min intestandbrowser); an apt outage is a red run at 15 min, re-run by a person. - From CannObserv/replicator#182, accepted 2026-11-08:
/etc/replicator/.envdestinations stay unscrubbed (REPLICATOR_BLOB_DIR,REPLICATOR_LOG_LEVEL,REPLICATOR_WHEELHOUSE_CREDENTIALS;REPLICATOR_REDIS_URL, which the recipe unsets;REPLICATOR_NOTIFY_*×6, whichtests/test_notify_failure.pyscrubs itself): they choose no code path, and no test reads them; corrected 2026-11-10:tests/test_inflight.py:113does readREPLICATOR_REDIS_URL, and prints it on failure (CannObserv/replicator#187);- the guard catches
boolsettings only, so a future path-choosingLiteral(asREPLICATOR_BLOB_BACKENDis) is added toPRODUCTION_ENVby hand; - a
Settings()validation failure in a loaded shell echoes a truncated copy of the env input into pytest output; - skipped:
-m browser, and a re-check of the 2026-11-03 occurrence (2 failed, names lost).
- From CannObserv/replicator#164, accepted 2026-11-09: the
disk-gccode change (+92/−8) was taken on the pinned digests and the upstream release, not reviewed here; the first timer run ofVERSION4 wasn't observed (the hand start stands in); the GC lists the SocratiCode plugin cache at1.14.0asinstalledwhile the npx pin is1.16.1(outside the child). - CannObserv/replicator#186 (2026-11-10): the cold-start half rests on a scratch 3 s delayed-launch run, not a cold host;
posix_fadviseeviction of the shell and driver didn't slow run 1. The test checks the sandbox's settled state, not how fast seccomp engages (6–55 ms measured). A process seen only before its argv reads--type=rendereris never checked. The 429 test no longer exercises the real clock;test_pacing.pyandtest_handler_pacing.pycover the arithmetic on fake clocks. - From CannObserv/replicator#183, accepted 2026-11-10: every
uv lock, and auv syncfrom a cold cache, needs the wrapper (uv run,uv lock --checkand a warm sync don't); the token reaches theuv_buildbackend uv spawns, and thewheelscommand'surllib(index host only); a forgotten deploy sync runs the old venv, as before; #136's real-wheel closure check skips unlessREPLICATOR_TEST_CLOSURE_WHEELSnames a fetched directory (CI 22 → 23 skips; locally the second skip); the CI bounds are unchanged on 2 runs of the new shape; no CI run with a cold uv cache. - From CannObserv/replicator#187, accepted 2026-11-10: the credential scrub covers
Settingsfields only (REPLICATOR_NOTIFY_*,REPLICATOR_CO_PYPI_CREDENTIALSand the repo.envPATs still reach a loaded test's environment; their own tests filter them); detection is by field name; the scrub is function-scoped (no import-, module- or session-scopeSettingstoday); theNOGROUPfail-open was shown by a unit test, not a scratch redis;-m integrationnot run locally. - From CannObserv/replicator#189, accepted 2026-11-12: nothing keeps the lifecycle rule, the IAM condition and the
tempconstant in step except the boot line and the provisioning record (docs/INFRASTRUCTURE.md); 7-day soft delete bills each reaped temp capture up to 7 more days; the two pre-split captures stay inco-gcs-blobsandco-gcs-replicator, metadata only; a capture and a fetched blob with the same digest are two objects. Archiver adopted a kept digest across kinds until CannObserv/archiver#499 (live 03:26:19Z); the first post-split capture hit exactly that case and got its own copy. - From CannObserv/replicator#188, accepted 2026-11-13:
fetch policy poll cycle recoveredis unit-tested only; its first live read-back is the next broker-unreachable window (Watching).after_failurescan exceed the cycle WARNINGs before it when a run includes stuck-cursor passes (ERROR-logged), unreachable through today's driver. No-m integrationrun (test_policy_integration.pyuntouched). - From CannObserv/replicator#201, accepted 2026-11-15:
-m integrationnot run (nothing undersrc/changed); the retiredtest_co_pypi_token.py's two real-loader cases (a mint from the named key with a writer key in ADC; a{}key through the real loader) are covered neither here nor upstream, offered on gregoryfoster/skills#652;tests/test_dependency_index.pyoverlapstest_co_pypi.py's index and lock checks, left alone; the ship's first deploy runsuv run --frozen --only-group co-pypiagainst the main checkout's dev venv (the wrapper's--projectis its own parent), a no-op there. - From CannObserv/replicator#202, accepted 2026-11-15:
RestrictNamespaces=stays off (it breaks Chromium's user-namespace sandbox, measured:No usable sandbox!; now pinned out by a test) andProtectControlGroups=stays off (Delegate=writes cgroupfs, CannObserv/observo#696);scripts/rehearse_reconnect.sh's scratch unit staysUser=exedev(it runs the dev checkout); the Playwright driver keeps the worker's environment (CR 11 of CannObserv/replicator#138); a session started before a group change keeps its boot-time groups (exe-init starts the VS Code server), bridged withsg replicator -c 'bash -c "…"'until a reboot; a Playwright bump reinstalls the shell asreplicatorfrom the live release's venv (DEPLOYMENT.md, The service user). - From CannObserv/replicator#206, accepted 2026-11-16:
-m integration,-m browserand-m gcsran locally oncae0dce, before CR 1 (096afd3, apyproject.tomlcomment only); CI re-ran browser and gcs on it. - From CannObserv/replicator#185, accepted 2026-11-16:
docs/INFRASTRUCTURE.mdis 10011 tokens exact against its 10000 budget (10101 before; the offline estimate 9925 reads under), left to a curation pass (CR 3, held); the Readers row is the bucket policy as read 2026-11-16, and nothing re-checks it; the deselected marks weren't run locally (docs and a docstring only).
Standing checks
df -h /at every check-in. The GC and its sensor own the disk since CannObserv/replicator#148: the monitorco-replicator-disk-headroompages at 85% and clears at 80%. No more hand reclaims; a page is read with the skill'sverifying.md. Last read 2026-11-10T21:1xZ: see Measured state. Before each patching run (gated at 85%), re-read it, and GCVERSION4 is in (CannObserv/replicator#164).The journal for: met 2026-11-11 (first production capture; Verification debt row discharged). Broker took part in the go-live live, per CannObserv/replicator#138's record [stated].stored a capture- Before any pin-move brief, and at each idle check: list the co-pypi index with no version bound (
$SCRATCH/idx.pyasco-pypi-reader@, or a lock probe at<1). Added 2026-11-13: the CannObserv/replicator#184 refreshes didn't, and missed 0.43.0 and 0.44.0. - Weekly since 2026-11-06 (operator); next due about 2026-11-21 (root 30 GiB since CannObserv/replicator#204: 84.0% of 29.42G usable is about 24.7G, so ~9.5G of headroom at 51.8%, measured 2026-11-15T04:1xZ): the
patching-hostsrun (knobposture scheduled,cadence weekly). The last full run was CannObserv/replicator#198 on 2026-11-14 (92 packages, 32 s down); before it, CannObserv/replicator#131 on 2026-10-06. The dry run needs--refresh-intotoo (gregoryfoster/skills#629), and the gate's real ceiling is 84.0% because the sensor rounds up. Best after the 05:15Z GC, which leaves the most headroom. Read the host memory before proposing it. Since CannObserv/replicator#148: the knob'squiet 05:15-06:15covers the GC's run, and the 85% line gates each step (gregoryfoster/skills#390). The run also covers the pinned browser's system libraries, whichlddchecks. - Idle check weekdays at 04:03 PT (operator, 2026-11-09; was every 12 h, rescheduled 2026-11-20): a recurring session cron
3 12 * * 1-5(UTC;3 11once PDT starts in March). The scheduler adds up to 30 min to a recurring job, so it fires by about 04:33 PT. It's read-only, a 5–10 line report, and re-creates itself at each run, because a recurring job expires after 7 days. A session restart drops it: re-create it from the saved prompt. Its issue query reads aboveLast issue seenas well as updated-since: the 2026-11-10 run missed CannObserv/replicator#185, filed before the previous report.
Watching
- Broker-unreachable windows. One on 2026-10-22, 04:13–04:16Z: 8 failed cycles, then recovered with no restart. One on 2026-11-11, 00:15:42–00:16:04Z, the co-broker reboot (CannObserv/broker#141): recovered, no restart; it surfaced CannObserv/replicator#188. The next window is the first live read-back of the tail's
fetch policy poll cycle recovered(shipped762fd1a): it should follow the tail'sfetch policy poll cycle failedWARNINGs. Threshold: a window that exhausts the boot budget, or recurs daily. Since 2026-10-24. - VS Code server builds. The live build is the client's own: 1.140.0 until at least 2026-11-03, 1.145.0 (
a717861) on 2026-11-06, then 1.145.2 (c4453f8) after the operator's window restart that evening. The CLI's agent host stages each new release about 6 h apart (1.145.1 at 06:33Z, 1.145.2 at 12:33Z on 11-06), and nothing starts them; GCVERSION2 removes them daily and they're fetched again. Paged once, 2026-10-28T17:00:24Z (85%, two staged), recovered by a hand-started GC run. The fix,VERSION4, is installed since 2026-11-09T00:27Z (CannObserv/replicator#164). The first report with a staged build should readkept vscode-servers … (<v>, the newest, newer than the live <v>). The firstVERSION4 'newest kept' line came on 2026-11-11's 05:27Z run (dcd3ad6, 1.145.3, newer than the live 1.145.2), as designed. Since then a third build (6904719) has been staged, about 660M each, and root is at 83% (22:47Z); the next run should dropdcd3ad6. Threshold: root ≥ 85%. - CI bounds after CannObserv/replicator#183 (the wheelhouse sync is gone): re-size once about 20 green
mainruns exist on the new shape; the rule already gives lint 5, test 20, browser 20, gcs 5 (ci.yml'sjobs:comment). Not filed. Since 2026-11-10. - Upstream from CannObserv/replicator#151: gregoryfoster/skills#649 (the service template starts Python through the
livelink) and gregoryfoster/skills#650 (hardlink identical files between releases: about one release, ~240M, back). The next vendor bump brings Wave 8 set D (55ede55f..8f2dbfae): no core change, three newtests/deployunit tests that pass on our units off-host (the 22:52Z note on CannObserv/replicator#151). Since 2026-11-14. - gregoryfoster/skills#652 (the co-pypi wheel fetcher, proposed from CannObserv/replicator#201) and gregoryfoster/skills#454 (the closure diff). When gregoryfoster/skills#652 ships, reinstall with
install.sh --co-pypi --replace; until then--checkcallsscripts/co_pypi_wheels.pyforeign. Since 2026-11-15. - The disk GC skipped 2026-11-15: CannObserv/replicator#202's reboot at 05:33:12Z came before the timer's randomized fire (05:15 + up to 30 min), and the timer is
Persistent=false(the shared GC's setting), so no catch-up. Last run 2026-11-14T05:27Z (197.5M); next 2026-11-16T05:18Z, the first after CannObserv/replicator#204's grow. Root 55%, so harmless. A reboot inside 05:15–05:45Z skips that day; the patching knob's quiet window already avoids it. Since 2026-11-15. - Docs near their 10000-token budget by the offline estimate (2026-11-16): TESTING 9992, the replicate issuer contract 9988, the outcome reference 9985, CONVENTIONS 9968, INFRASTRUCTURE 9925 (exact 10011, over), STREAMS 9902. The exact count runs higher than the estimate here (INFRASTRUCTURE at 2.52 bytes/token against the repo's 2.65), so others may be over too. A curating-context pass owns it; no gate fails today.
prmerge mode since 2026-11-17 (CannObserv/replicator#208): two standing bot PRs, fromskills/submodule-refresh(the refresh hook) andcontext/weekly-measurement(the cadence), each merged by an operator with a merge commit; until merged, the vendored skills don't advance and the weekly row isn't onmain. Read their state at each check-in. A rollback target is now a merge commit (docs/DEPLOYMENT.md). Threshold: either PR open more than 7 days, or CI red on it. Since 2026-11-17.- gregoryfoster/skills#675 (the cadence template's
actions/checkoutSHA pin) and gregoryfoster/skills#674 (the knob-flipping branch carries its own re-render), both from CannObserv/replicator#208. When gregoryfoster/skills#675 ships and is vendored, re-runinstall-cadence.shfrom the main checkout, through a PR, and check--check'scheckout pin:line. Since 2026-11-17.
Measured state
-
Production runs a release as
replicator(CannObserv/replicator#202):/srv/replicator/live→releases/59fb24082f7e(deployed 2026-11-15T23:44:13Z), previousf2a43335c6b0; both unitsNoNewPrivileges,ProtectSystem=strict,ProtectHome,PrivateTmp; NRestarts 0, pid cwd the release, nothing mapped from/home.capture,browser_fetchandpersistareenabled. co-core 0.48.2 ×3.mainis atd549a06, clean, atorigin/main; live trails it by docs and a docstring (CannObserv/replicator#185, not deployed). Only INFO since the start. Deploys:scripts/deploy.sh(core version 1,KEEP=0), also viashipping-work; the co-pypi wrapper is the shared one (version 1). Measured 2026-11-16T20:4xZ. -
Fetch policy: the 2026-11-15T23:44Z start's replay read 503 messages,
tracked_hosts10.without a generationnever logged since 2026-10-22. -
Browser:
REPLICATOR_BROWSER_FETCH_ENABLED=true. Norender="browser"command yet. Captures: 3 stored; the two pre-split (2026-11-11) are inco-gcs-blobsandco-gcs-replicator; since 02:41Z they go toco-gcs-captures(temp/, kept underblobs/), the first01M76WR77JYHK5QFS31W3JXQ44at 03:27:48Z. -
Disk: root 30 GiB since CannObserv/replicator#204: 55% (
df -h) at 2026-11-16T15:47Z; two releases at 248M each; sensoralert_at85,clear_at80; the GC ran 2026-11-16T05:18Z, exit 0, 166.0M removed, 0 failed. -
co-core: 0.48.2 ×3 from the co-pypi index (Path C, CannObserv/replicator#183), pin
>=0.48.0,<0.49(CannObserv/replicator#206), locked with URLs and sha256. The index's newest is 0.48.2 ×3 [measured 2026-11-15T23:5xZ].uv lock --checkclean.uv audit --frozenclean, 0 open Dependabot alerts (2026-11-03). -
CI: four jobs bounded (CannObserv/replicator#180); no wheelhouse sync since CannObserv/replicator#183;
uv syncgets a masked WIF access token.main's tipd549a06: all four green. uv: host 0.13.5, CI pinned to>=0.13.0,<0.14(0.13.8), theuv_buildbound's range; nouv syncwarns since CannObserv/replicator#195 (d5926ab, run 43512090888). -
Context, exact: POLITENESS 8106; outcome reference 9985 of 10000, CONVENTIONS 9968, STREAMS 9875, COMMANDS 9840, the issuer contract 9839, ENVIRONMENT 9209, the issuer reference 8465, AGENTS.md 5969 of 6000; after CannObserv/replicator#183, AGENTS.md 5897 and DEPLOYMENT 8014 [stated by the implementer].
-
Host addressing: eth0
10.42.0.42/16; no IPv6 default route. -
Env loading:
scripts/load_env.shis the single entry point. -
Tests:
PRODUCTION_ENVscrubs everyboolsetting, guarded (CannObserv/replicator#182); everySettingscredential too (CannObserv/replicator#187), so a loaded shell matches a clean one with the live broker URL set. Baseline ond549a06: 2062 passed, 1 skipped (test_co_core_closure_diff.py:428;test_deploy.py:895runs and passes from this post-boot session, which counts as an exe.dev session), 105 deselected;tests/ci44;tests/deploy32 on the host;-m browser31 [the implementer's, on7b338ff].
Overlaps
| Surface | Items |
|---|---|
the fetch outcome reference (9985), docs/CONVENTIONS.md (9968), STREAMS (9875), the issuer contract (9839) |
any new row needs a curating-context pass first |
AGENTS.md, docs/DEPLOYMENT.md, the manifest |
none open; AGENTS.md has 59 tokens of headroom since CannObserv/replicator#202 [stated] |
docs/POLITENESS.md, src/worker/policy.py |
CannObserv/replicator#163 B (parked; A shipped 7460a5c) |
pyproject.toml, uv.lock, docs/DEPENDENCIES.md, tests/test_dependency_index.py |
the next pin move; CannObserv/replicator#169 (parked) |
.github/ (ci.yml, dependabot.yml, tests/ci guards; Dependabot action PRs land here too) |
CannObserv/replicator#169 (parked) |
skills-vendor/ submodule (the hook bumps it on its own, untested: 4cac2a2 pushed red; later bumps green) |
the next bump (Wave 8 set D, re-copy tests/deploy), the gregoryfoster/skills#536 adoption |
docs/SKILLS.md (about 9314) |
the gregoryfoster/skills#536 adoption (:58) |
src/core/bus_client.py (build_bus_client) |
CannObserv/replicator#178 (no change unless broker withholds the grant) |
docs/INFRASTRUCTURE.md (the production temp store table) |
a curation pass: 10011 exact of 10000 after CannObserv/replicator#185 (Watching) |
src/worker/policy.py |
CannObserv/replicator#188, CannObserv/replicator#163 B (parked) |
scripts/deploy_hooks.sh, deploy/ units |
CannObserv/replicator#203 (parked on gregoryfoster/skills#641; its adoption adds units and inputs) |
Log
- 2026-10-24: created. 8 open: 5 in tiers, plus 1 new child (CannObserv/replicator#153); 3 parked (CannObserv/replicator#64, CannObserv/replicator#135 and its CannObserv/replicator#144); 0 closures proposed. Reworks: CannObserv/replicator#63 (selection settled), CannObserv/replicator#76 (cohort shape). CannObserv/provisioner#61 closed today, so CannObserv/replicator#148 is half unblocked. Gate green on
42b728c. Next: CannObserv/replicator#153. - 2026-10-24: filed CannObserv/replicator#153 (co-core catch-up; operator-approved) and posted its brief. CannObserv/replicator#148 interim reclaim: 81% → 77% (SocratiCode 1.14.0
_npxand Claude Code 2.1.301;CachedExtensionVSIXskept, because the live VS Code server holds a handle in it). Replies posted on CannObserv/archiver#377 and CannObserv/cannobserv#495 (operator consent). Next: CannObserv/replicator#153. - 2026-10-25: check-in 1: CannObserv/replicator#153 verified, deployed and closed (PR CannObserv/replicator#155, the only new number). The baseline is 1468/1, with the skip identical. My
[extract]claim was corrected in place on CannObserv/replicator#153's body and brief. Harvest: the For #138 notes are in CannObserv/replicator#138's brief. Blockers: CannObserv/provisioner#85 closed (CannObserv/replicator#148's line rewritten); gregoryfoster/skills#425, gregoryfoster/skills#469, CannObserv/cannobserv#554, CannObserv/watcher#374 and CannObserv/archiver#366 still open. Found: the vendoredmcp-driver.mjshas noupdatesubcommand; filing upstream is pending the operator. Next: CannObserv/replicator#138. - 2026-10-25: filed gregoryfoster/skills#483 (operator):
mcp-driver.mjsneeds anupdatesubcommand, so a cappedcodebase_updatecan run outside a session. Next: CannObserv/replicator#138. - 2026-10-26: check-in 2: CannObserv/replicator#138 verified, deployed and closed (PRs CannObserv/replicator#157 and CannObserv/replicator#158, 49 commits); capture is live and the debt is listed. New: CannObserv/replicator#156 and CannObserv/replicator#159 (bugs, Tier 1) and CannObserv/replicator#160 (filed, Tier 3; red Dependabot security-update runs since 09:15Z). gregoryfoster/skills#425 closed, so CannObserv/replicator#148 is unblocked once the skill is vendored. CannObserv/archiver#366 is deployed dark; F1 is still open. Corrected in place: CannObserv/replicator#159 isn't live in production (no sweeper under the object store). Harvest notes are on CannObserv/replicator#63, CannObserv/replicator#76, CannObserv/replicator#148 and CannObserv/replicator#156. Next: CannObserv/replicator#159.
- 2026-10-26: check-in 3: CannObserv/replicator#159 verified, deployed and closed (
9b37afe..66321b5, 5 commits, fast-forward); the CI flake is fixed; the baseline is 1691/1, with the skip identical. No new numbers. Harvest: #159's limits and its unbounded-grace note went under Accepted limitations, with a re-entry trigger. Blockers: CannObserv/watcher#374 part 1 shipped (Watcher's producer stopped at 18:55Z), and F1 is still open; CannObserv/provisioner#85 confirmed closed. Pre-flight for CannObserv/replicator#148 found no co-status tenant or ACL rule, so CannObserv/status#67 was filed (operator), with the edge in the same request, the CannObserv/notifier#70 way. Root 80%. Next: CannObserv/replicator#148. - 2026-10-27: check-in 4: CannObserv/replicator#148 verified and closed (
66321b5..6010a66; the GC and sensor are live, the monitor is enabled, the rehearsal was confirmed). CannObserv/status#67 closed. No new numbers. Since check-in 3, this thread also coordinates CannObserv/replicator#135 (operator, 2026-10-26): the read-back is posted, CannObserv/replicator#144 is unparked, and the Parked rows are reworded. The #148 brief was revised after the Status agreement (a disabled monitor still pages; the hook vendored the skill on its own). Pre-flight for CannObserv/replicator#156 found the same gap in Watcher's copy of the guard; filed after the fix (operator). Watching: the daily 76→81% re-stage (gregoryfoster/skills#504). Baseline 1691/1. Next: CannObserv/replicator#156, then CannObserv/replicator#144. - 2026-10-27: check-in 5: CannObserv/replicator#156 verified and closed (
6010a66..6efad10, 6 commits; the guard probed live in the service venv). Filed CannObserv/watcher#402 (the port), and CannObserv/replicator#161 and CannObserv/replicator#162 into Tier 3 (operator). CannObserv/broker#116: Replicator's half read back clean, and that debt row is discharged. Corrected: check-in 2's doc counts were estimates; COMMANDS.md is 10096 exact and unchanged since9b37afe. The CannObserv/replicator#144 pre-flight was re-run and its brief posted: the unordered path stays (CannObserv/cannobserv#579 open), and the manifest needs editing too. Baseline 1735/1. Next: CannObserv/replicator#144. - 2026-10-28: check-in 6: CannObserv/replicator#144 verified and closed (
6efad10..02aa79f, after the bot'saa4c697; live, replay clean, no F1-as-future wording). That emptied CannObserv/replicator#135, which closed with its summary, CannObserv/watcher#368 notified, and its process log filed as gregoryfoster/skills#513. Filed CannObserv/replicator#163 (two triggers, parked). Corrected in place: the CannObserv/replicator#144 brief's COMMANDS.md pointer, and the coordinator's own claim 'noted on gregoryfoster/skills#483' (nothing was posted; posting waits for the operator). Tier 3 order decided (operator). The CannObserv/replicator#160 pre-flight foundvirtualenv21.15.1 movingpackagingtoo, and the Dependabot runs stopped. CannObserv/replicator#63 is still blocked (no render-field issue filed). Baseline 1735/1. Next: CannObserv/replicator#160. - 2026-10-28: check-in 7: CannObserv/replicator#160 verified and closed (
02aa79f..f3db89b, lock only;uv auditclean; no restart). Corrected in place on CannObserv/replicator#160: the brief's 21.14.x option and the body's 21.14.9 floor. Operator asks done: posted on gregoryfoster/skills#483 (one line reference corrected in place), and filed CannObserv/cannobserv#602 (therenderfield), which now blocks CannObserv/replicator#63. Disk paged at 17:00Z (85%: a third VS Code build); decided GC now plus a VS Code reload (operator). Pre-flight for CannObserv/replicator#76 found two cohort hazards (setup-uv'sprune-cachedefault on a public repo,ubuntu-latest→ 26.04); brief posted. Baseline 1735/1. Next: CannObserv/replicator#76. - 2026-10-28: the disk page resolved. The operator rebooted the VM; the worker came back clean on
f3db89b, with every cgroup and limit as before. The reboot didn't switch VS Code builds: the client is on 1.140.0, and the staged builds are the agent host's unused downloads (corrected the coordinator's own pre-reboot claim, in memory and here). GC started by hand (operator-authorized): 1.4G removed, 73%, recovered 22:20:25Z. Measurement and a VERSION 3 concern posted on gregoryfoster/skills#504. New: CannObserv/replicator#164 (GC VERSION 3), slotted in Tier 3 and held on the VS Code client's version. Next: CannObserv/replicator#76. - 2026-10-28: decided (operator): CannObserv/replicator#164 waits for the operator's VS Code client update, then runs after CannObserv/replicator#76. Next: CannObserv/replicator#76.
- 2026-10-30: check-in 8: CannObserv/replicator#76 verified and closed (
f3db89b..cda2861; Dependabot proposed and landed three action bumps, all green; OIDC minted on its runs). The uv half was re-filed as CannObserv/replicator#169 and parked on CannObserv/cannobserv#554. CannObserv/replicator#165 (new) was closed as met by the hook'sb520e7f, its preflight passing (operator). Every child the exit criterion names is now closed and verified; the remaining unblocked children are CannObserv/replicator#162, CannObserv/replicator#161 and CannObserv/replicator#164 (held on the VS Code client). Corrected in place: CannObserv/replicator#162's body (a generated workflow can't carry a per-doc check). Baseline 1756/0. Next: CannObserv/replicator#162. - 2026-10-30: check-in 9: CannObserv/replicator#162 verified and closed (
cda2861..ec96147, after the hook's red4cac2a2; CI green by job, exact counts match, suite 1756/0, re-embed done by the watcher). New: CannObserv/replicator#170 and CannObserv/replicator#171, both upstream follow-ups. CannObserv/replicator#171's steps had already landed, so it was re-scoped to its drift test. Decided (operator): CannObserv/replicator#170 → CannObserv/replicator#171 → CannObserv/replicator#161. Two upstream asks drafted, not filed. CannObserv/replicator#164 still waits on the VS Code client. Next: CannObserv/replicator#170. - 2026-10-30: check-in 10: CannObserv/replicator#170 verified and closed (
ec96147..ccd034f, 7 commits; CI green by job, 1783/0, AGENTS.md 5960,rcli PINGfrom a clean shell). Nothing new. Decided (operator): CannObserv/replicator#161 takes all 12 ranges, with one doc clause. Next: CannObserv/replicator#171. - 2026-10-30: check-in 11: CannObserv/replicator#171 verified and closed (
ccd034f..be3d735; CI green by job, 1793/0, COMMANDS 9800, re-embed by the watcher). co-core v0.40.0 shipped, so CannObserv/replicator#172 was filed into Tier 2 (operator). gregoryfoster/skills#469 closed as a duplicate; CannObserv/replicator#151 now waits on gregoryfoster/skills#372 and gregoryfoster/skills#345. CannObserv/cannobserv#602 was decided and briefed (v0.40.1, behind CannObserv/cannobserv#609). Declined (operator): the uncapped-launch guard and the three upstream asks. Next: CannObserv/replicator#161, then CannObserv/replicator#172. - 2026-10-31: check-in 12: CannObserv/replicator#161 verified and closed (
be3d735..37b6eca; CI green by job, 1819/0, live since 16:00Z; all 12 rows and embedded forms probed refused in the service venv; exact budgets match). The SOCKS-path test gap closes by construction (oneblocked_networkssource), so the verify recipe now separates rows from forms. Filed (operator): CannObserv/watcher#412 (the 12 rows) and CannObserv/replicator#173 (rawssl.SSLError, Tier 1, after CannObserv/replicator#172);100:0:0:1::/64left out. CannObserv/cannobserv#602 cleared to build; no 0.40.1 yet. Next: CannObserv/replicator#172, briefed. - 2026-11-01: check-in 13: CannObserv/replicator#172 verified and closed (
37b6eca..2a6e0f2; CI green by job, 1819/0, co-core 0.40.0 ×3 live, replay clean; the re-embed was already done by the coordinator session's watcher). New: CannObserv/replicator#174 (unifiedshipping-work, Tier 3, and now a prerequisite of CannObserv/replicator#151 via gregoryfoster/skills#345) and CannObserv/replicator#175 (a stale integration test, Tier 1). v0.40.1 is on the index, so CannObserv/replicator#63 is unblocked; CannObserv/watcher#414 waits on it.uv auditfound 4pyasn1advisories, folded into CannObserv/replicator#173. Decided (operator): CannObserv/replicator#174 → CannObserv/replicator#173 → CannObserv/replicator#175 → CannObserv/replicator#63. Next: CannObserv/replicator#174, briefed. - 2026-11-02: check-in 14: CannObserv/replicator#174 verified and closed (
2a6e0f2..337623a: the hook's bump, 3 commits, then the hook's second bump; CI green by job on both; 1819/0;rcli PINGfrom a clean shell; no restart). My brief's "breaks at release 2" was corrected in place on the issue. No new numbers. Harvest: aFor #151note, the gregoryfoster/skills#536 adoption added to Waits, and #174's limits. Blockers unchanged: gregoryfoster/skills#345 is done on a branch, not merged. Next: CannObserv/replicator#173, briefed. - 2026-11-03: check-in 15: CannObserv/replicator#173 verified and closed (
337623a..0a3c5af; CI green by job; live on1029e6a, replay clean;pyasn10.6.5, audit clean, alerts #23–#26 fixed). New: CannObserv/replicator#176 (Dependabot setup-uv 10.3.3), merged by the coordinator as64c92c0on the operator's decision. Harvest: the OOM integration flake into CannObserv/replicator#175's brief (traced), For #63 notes, #173's limits; an unidentified default-suite flake (1 of 10 runs) added to Watching. Next: CannObserv/replicator#175, briefed. - 2026-11-03: check-in 16: CannObserv/replicator#175 verified and closed (
64c92c0..81396a4; CI green by job;-m integration57/57 twice on a scratch broker; no restart). CONVENTIONS is at 9968 exact. Decided (operator): split the 0.40.1 floor from CannObserv/replicator#63, filed as CannObserv/replicator#177 (Tier 2, next). Harvest: For #63 notes, #175's limit. Blockers unchanged: CannObserv/replicator#151 (upstream), CannObserv/replicator#164 (the VS Code client is still 1.140.0). Next: CannObserv/replicator#177, briefed; then CannObserv/replicator#63, the last unblocked child. - 2026-11-03: check-in 17: CannObserv/replicator#177 verified and closed (
81396a4..93fb513; CI green by job; live on93fb513, 0.40.1 ×3, replay clean; re-embed confirmed). My brief's decode site was corrected in place, twice. New: CannObserv/replicator#178 (broker'sCLIENT SETINFOquestion), answered on the operator's decision: take the grant. Next: CannObserv/replicator#63, the last unblocked child, briefed: design doc first, and announcement replies posted by the coordinator with consent. - 2026-11-06: check-in 18: CannObserv/replicator#63 verified and closed (
93fb513..0de3ce5; CI green by job; default 1902/0, browser 31, integration 57 re-run by the coordinator; live on0de3ce5withbrowser_fetch: enabled). With the operator's go-ahead: announced on CannObserv/watcher#414 and CannObserv/archiver#379, filed CannObserv/cannobserv#635 and CannObserv/replicator#179 (co-core 0.41, from the cannobserv coordinator's epic comment). Re-scan: CannObserv/replicator#163 un-parked (trigger A met: CannObserv/broker#95 closed); CannObserv/replicator#169 re-read (CannObserv/cannobserv#554 closed, CannObserv/cannobserv#615 open), still parked. Next: CannObserv/replicator#179, briefed; then CannObserv/replicator#163 A. - 2026-11-06: check-in 19: CannObserv/replicator#179 verified (
53918a7..74c69a7; CI green by job; default 1907/0, browser 31, integration 57 re-run by the coordinator; live on74c69a7, 0.41.0 ×3). New: CannObserv/replicator#180 (CI timeouts, Tier 3). Operator: patching cadence weekly (knobcadence weekly,beb6d3b);.env.bak-63deleted; CannObserv/replicator#164 rescoped to GCVERSION4, waiting on gregoryfoster/skills#551. gregoryfoster/skills#345 vendored, so CannObserv/replicator#151 waits on gregoryfoster/skills#372 alone. Next: CannObserv/replicator#163 A, briefed; then CannObserv/replicator#180. - 2026-11-06: check-in 20: CannObserv/replicator#163 trigger A verified (
beb6d3b..7460a5c; CI green by job; default 1907/0, boundaries 97 re-run by the coordinator; no restart); B parked again with the report's notes. No new numbers. With the operator's go-ahead: nudged gregoryfoster/skills#551, filed CannObserv/cannobserv#640. VS Code client now 1.145.2. Next: CannObserv/replicator#180, briefed, the last unblocked child. - 2026-11-07: check-in 21: CannObserv/replicator#180 verified (
7460a5c..7b21e86; CI green by job;tests/ci27, default 1914 re-run by the coordinator; no restart). The watched flake recurred once (5 failed, names lost), 14 clean after. Filed CannObserv/replicator#181 (operator). CannObserv/replicator#169 re-read: Watcher is on the co-pypi index, Replicator's reader identity can read it; un-parking part 1 proposed. Blockers unchanged; gregoryfoster/skills#551 scheduled upstream. Next: CannObserv/replicator#181. - 2026-11-08: check-in 22: CannObserv/replicator#181 verified (
7b21e86..83af705; CI green by job;tests/ci37, default 1913/1 in a clean shell; no restart). The watched flake solved:PRODUCTION_ENVmisses the capture and browser-fetch flags, sopytestafterload_env.shfails 5 tests every time; my #180 verification corrected in place. Filed CannObserv/replicator#182 (bug, next) and CannObserv/replicator#183 (#169 part 1, the co-pypi index) on the operator's decision. Next: CannObserv/replicator#182. - 2026-11-08: check-in 23: CannObserv/replicator#182 verified (
83af705..86c8986; CI green by job; default 1914/1 in a clean shell and afterload_env.sh,tests/ci37, ruff clean; no restart); the Verify recipe's clean-shell caveat is retired. No new numbers. gregoryfoster/skills#551 shipped and was vendored (edb63bc), so CannObserv/replicator#164 is unblocked; pre-flight planupdate2 → 4, read-only. Decided (operator): CannObserv/replicator#164 → CannObserv/replicator#183. Next: CannObserv/replicator#164, briefed. - 2026-11-09: check-in 24: CannObserv/replicator#164 verified (host-only: GC
VERSION4 at 00:27Z, re-plancurrent, timers enabled, first run exit 0, root 74%; no repo change, no restart). co-core 0.42.0 found on the index; filed CannObserv/replicator#184 after CannObserv/replicator#183 (operator). Pre-flight for CannObserv/replicator#183 found the reader-key hazard (ADC here is the writer) and 9 more wheelhouse sites; brief posted. Next: CannObserv/replicator#183. - 2026-11-10: check-in 25 (operator: file and pre-flight the browser CI failure):
browserred onb5d4f2f, the hook's submodule bump; reproduced 3 in 30 locally, traced to the sandbox test's poll; filed CannObserv/replicator#186, ordered before CannObserv/replicator#183 (operator). Found CannObserv/replicator#185 (docs, from Watcher, filed 2026-11-09), waiting on CannObserv/watcher#420. Idle checks every 12 h set up 2026-11-09. Next: CannObserv/replicator#186. - 2026-11-10: check-in 26: CannObserv/replicator#186 verified and closed (
b5d4f2f..c74be97, 4 test commits, fast-forward; step 1 showed a test race, not a sandbox gap; scope extended to the 429 pacer flake (operator)); CI green by job onc74be97; coordinator re-runs 15/15,-m browser31, 1914/1; no restart. No new numbers. Harvest: the For #183 note and a pre-flight refresh on CannObserv/replicator#183; #186's limits under Accepted limitations. Blockers unchanged (gregoryfoster/skills#372, CannObserv/broker#85, CannObserv/cannobserv#615, CannObserv/watcher#420: no new comments). Next: CannObserv/replicator#183, briefed. - 2026-11-10: check-in 27: CannObserv/replicator#183 verified and closed (
c74be97..294e489, merged asbc4c5ec; co-core from the co-pypi index, the wheelhouse retired, reader key renamedREPLICATOR_CO_PYPI_CREDENTIALS); CI green by job; live since 18:36:53Z, 0.41.0 ×3, replay clean, 12 blobs stored; coordinator 1959/2 clean and loaded (minus the broker URL). New: CannObserv/replicator#187 (a loaded-shell test prints the broker password), Tier 1, next (operator), briefed. Harvest: For #184 and a pre-flight on CannObserv/replicator#184; #183's limits; #76'sbrowserassert limit settled; #182's 'no test reads them' corrected in place; CannObserv/replicator#169 re-read (one of three conditions met). Blockers unchanged. Next: CannObserv/replicator#187. - 2026-11-10: check-in 28: CannObserv/replicator#187 verified and closed (
294e489..10d9ef3; credential settings scrubbed and guarded, no URL in a diff,inflight.pyreads its file alone); CI green by job; coordinator 1961/2 clean and loaded with the live URL (0 password hits); no restart. Corrected in place: my brief's 'a stale shell value fails closed' (it could open the patching gate; the implementer found it). No new numbers. Archiver's capture issuance is starting (CannObserv/replicator#138): the capture debt row gains its discharge path; the stale capture-contract Status line folds into CannObserv/replicator#184 (operator). Blockers unchanged. Next: CannObserv/replicator#184. - 2026-11-11: check-in 29 (operator: pre-flight CannObserv/replicator#189): Archiver's D4 asks for a captures-only bucket; pre-flight and brief posted (design doc first; one bucket's two tiers conflict on soft delete,
updateand the lifecycle scope; persist routes by source; replicate will refuse captures). Also slotted CannObserv/replicator#188 (from the 00:15Z co-broker reboot). The capture debt row is discharged (first production capture 14:46:26Z, closed with Archiver on CannObserv/replicator#138). Order (operator): CannObserv/replicator#189 → CannObserv/replicator#184 → CannObserv/replicator#188. Root 83% (three VS Code builds). Next: CannObserv/replicator#189. - 2026-11-12: check-in 30: CannObserv/replicator#189 verified and closed (
10d9ef3..6e0b19f, PR CannObserv/replicator#190 merged asba37286; captures inco-gcs-captures, provisioned by the operator, object-level grant probe PASS); CI green by job; live since 02:41:35Z; the first post-split capture read back intemp/and kept inblobs/, raw buckets untouched; coordinator 2001/2 clean and loaded. Reply posted on CannObserv/archiver#380. CannObserv/replicator#184's fold-in voided (done by #189) and its pre-flight refreshed. Next: CannObserv/replicator#184. - 2026-11-13: check-in 31: CannObserv/replicator#184 verified and closed as 0.41 → 0.44 (
6e0b19f..6eb65b8, PR CannObserv/replicator#191; widened at review because 0.43.0 and 0.44.0 were already on the index); CI green by job; live since 2026-11-12T23:18:59Z on 0.44.0 ×3, replay clean; coordinator 2001/2 clean and loaded. Corrected in place: my stale index claim on CannObserv/replicator#184 (retitled); a standing check now lists the index unbounded. CannObserv/replicator#188 briefed (a second silent recovery path atpolicy.py:575). Root 78%. Next: CannObserv/replicator#188. - 2026-11-13: idle check found co-core 0.45.0 ×3 on the index (tagged 05:36Z). Filed CannObserv/replicator#192 (operator), closure diff version-only [measured]; slotted after CannObserv/replicator#188. Nothing else moved: build
6eb65b8, NRestarts 0, 0 WARNING/ERROR; blockers unchanged; root 81% (extension update). Next: CannObserv/replicator#188. - 2026-11-13: check-in 32: CannObserv/replicator#188 verified and closed (
6eb65b8..762fd1a, PR CannObserv/replicator#193, the only new number); CI green by job; live since 19:00:28Z, replay 501/10 clean; coordinator 2005/2 clean and loaded. The recovery line is unit-tested only (Accepted limitations, Watching). Filed CannObserv/processor#134 and CannObserv/watcher#436 (operator): same gap, traced. CannObserv/replicator#192 pre-flighted: index still 0.45.0;test_dependency_index.py's>=0.45bound added to scope. Blockers unchanged. Next: CannObserv/replicator#192. - 2026-11-13: check-in 33: CannObserv/replicator#192 verified and closed (
762fd1a..da5a19d, PR CannObserv/replicator#194); CI green by job; live since 20:48:46Z on 0.45.0 ×3, replay 503/10 clean; coordinator 2005/2 clean and loaded; index still 0.45.0. New: CannObserv/replicator#195 (filed by the implementer, operator's call), slotted Tier 3 and pre-flighted; it's the last unblocked child. Blockers unchanged. Next: CannObserv/replicator#195. - 2026-11-13: check-in 34: CannObserv/replicator#195 verified and closed (
da5a19d..d5926ab, PR CannObserv/replicator#196; CI's uv pinned to the bound's range, warning gone in CI and on the host; no restart); coordinator 2008/2 clean and loaded,tests/ci44. New: CannObserv/replicator#197 (archiver's patch-window notice), closed as no impact with the operator's consent, correction on CannObserv/archiver#375. Blockers re-read: CannObserv/cannobserv#640 met (v0.42.0); the rest unchanged. No child unblocked; the hand-off is the blockers. - 2026-11-14: check-in 35: the weekly patching run CannObserv/replicator#198 verified and closed (92 packages, reboot 00:50:45Z, 32 s down, 0 pending after);
3d16e77(knob exception) and the hook'sa46e650, CI green by job on both; coordinator 2009/1,tests/ci44. gregoryfoster/skills#372 pre-flighted upstream (a plan first), still open. Still no child unblocked: the hand-off is the blockers; first to clear is CannObserv/watcher#420 → CannObserv/replicator#185, no earlier than 2026-11-16. - 2026-11-14: check-in 36: nothing merged since
a46e650; the 13:57Z idle check found co-core 0.46.0 and 0.46.1 ×3 on the index, and gregoryfoster/skills#372 Part A shipped (f728abd5), unblocking CannObserv/replicator#151. Filed CannObserv/replicator#199 (operator), closure diff version-only [measured]. Decided (operator): CannObserv/replicator#199 → CannObserv/replicator#151. Coordinator note on CannObserv/replicator#151: the vendored pin4f26e316lacks the skill;KEEPsized against 3.6G free. Next: CannObserv/replicator#199. - 2026-11-14: check-in 37: CannObserv/replicator#199 verified and closed (
a46e650..91e87dc, PR CannObserv/replicator#200); CI green by job; live since 19:01:13Z on 0.46.1 ×3, replay 509/10 clean; coordinator 2009/1 clean and loaded,tests/ci44. My brief'swp_view_*_urlclaim corrected in place (not in Replicator's wheels). Decided (operator): keep the closure diff; the shared co-pypi wrapper is CannObserv/replicator#201 (filed), after CannObserv/replicator#151, its fetcher built upstream-ready. Next: CannObserv/replicator#151, briefed. - 2026-11-14: check-in 38: CannObserv/replicator#151 verified and closed (
91e87dc..b61703a; CI green by job onMandC): replicator runs a root-owned release under/srv/replicator, rollback drilled, coordinator 2028/1 andtests/deploy32; myKEEParithmetic corrected in place; agent memory retired. Root is 83.72%: decided (operator) to grow the disk, filed CannObserv/replicator#204 into Tier 1. New follow-ups CannObserv/replicator#202 and CannObserv/replicator#203 into Tier 3. For #201 note posted. Next: CannObserv/replicator#204. - 2026-11-15: check-in 39: CannObserv/replicator#204 verified (the operator's grow to 30 GiB, reboot 23:55:50Z, no root commands needed;
b61703a..7706bf0docs, deployed 00:13Z; CI green by job on7706bf0); root 54.46%, the worker on7706bf05cc16, coordinator 2028/1,tests/deploy32. The sysctl drop-in's drift discharged (operator-authorized install). Patching has ~8.9G headroom. Next: CannObserv/replicator#201, pre-flighted. - 2026-11-15: check-in 40: CannObserv/replicator#201 verified (
7706bf0..a8900f5, PR CannObserv/replicator#205; CI green by job ona8900f5; livea8900f5d15a2, the first production mint through the shared wrapper; coordinator 2044/1, the new fetcher's 12 wheels and the dry pin move reproduced). The manifest re-embed debt discharged. Upstream: gregoryfoster/skills#652 filed, gregoryfoster/skills#454 commented (operator). Found: the worker's uid has passwordless sudo, unsandboxed. Decided (operator): CannObserv/replicator#202's shared-group model; CannObserv/replicator#202 next, briefed. - 2026-11-15: check-in 41: CannObserv/replicator#202 verified (
a8900f5..f2a4333; CI green by job; the worker runs asreplicator, confined, after the 05:33Z reboot; a production capture after the switch; coordinator 2062/1, host deploy tests 181). Filed CannObserv/replicator#206 (operator): co-core 0.46 → 0.48, widened to 0.48.2 at pre-flight; closure diff Drive-only. The GC skipped 2026-11-15 (the reboot,Persistent=false). Next: CannObserv/replicator#206. - 2026-11-16: check-in 42: CannObserv/replicator#206 verified (
f2a4333..59fb240; CI green by job; live on 0.48.2 ×3 asreplicator; coordinator 2062/1 and 181 counting one session-premise test run in the foreground; the code index caught up). CannObserv/watcher#420 closed 2026-11-15 (gate waived), so CannObserv/replicator#185 is unblocked; pre-flight posted (two more stale sites; the Reader row needs the operator's policy read). New: #207 (the PR, closed). Next: CannObserv/replicator#185. - 2026-11-16: check-in 42, continued: the operator chose to park CannObserv/replicator#203 on gregoryfoster/skills#641; the scope ask is posted there and the park note on CannObserv/replicator#203. CannObserv/replicator#185 is the last unblocked child.
- 2026-11-16: check-in 43: CannObserv/replicator#185 verified (
59fb240..d549a06; CI green by job; docs and a docstring; Readers from the operator's live policy; coordinator 2062/1; not deployed, by the operator). The disk GC ran 05:18Z (exit 0). The exit criterion is met; no child is unblocked. - 2026-11-16: check-in 44: the operator chose parking-lot mode. Retitled; every parked row rewritten brief-grade and re-read (co-core 0.48.2 still decodes
generationas optional; Archiver still on 0.38; CannObserv/archiver#408 and CannObserv/cannobserv#615 unchanged). - 2026-11-16: the run journaled at the park (operator's approval): process-log entry filed upstream as gregoryfoster/skills#668.
- 2026-11-17: check-in 45: the operator asked for PR mode; filed and pre-flighted CannObserv/replicator#208. Idle check 08:43Z normal: the hook's
85f8d7bgreen by job; live59fb24082f7e, NRestarts 0; co-core 0.48.2 still newest. - 2026-11-17: check-in 46: CannObserv/replicator#208 verified (
85f8d7b..a072ba5, the first GitHub merge commit; push CI green by job); not deployed (operator). Two debt rows (firstpr-mode cadence and hook runs), one Watching row (the standing bot PRs). CR 6 asked. - 2026-11-17: CR 6 (operator's answer): filed upstream as gregoryfoster/skills#675.
- 2026-11-18: check-in 47: merged the cadence PR #210 (operator's request; close/reopen for CI, green by job, merge commit
796186f); the firstpr-mode cadence row discharged. Filed and briefed CannObserv/replicator#211 (co-core 0.48 → 0.50) out of the parking lot; parked rows re-read against 0.50.0: CannObserv/replicator#163 B unchanged. - 2026-11-18: check-in 48: CannObserv/replicator#211 verified and live (
753ecb9609a0, co-core 0.50.0 ×3, the first GitHub-merge-commit deploy); the hook's firstpr-mode run opened PR #212 (debt discharged); two brief claims corrected in place. - 2026-11-18: merged the refresh PR #212 (operator's request): skills
063ba5d→817963f, merge commit64a2c7a, itspushCI green by job; main checkout pulled and the submodule checked out to the pointer, doctor exit 0, clean. Not deployed: no shipped code changed. - 2026-11-20: idle check rescheduled to weekdays 04:03 PT (operator); status read 17:59Z: two one-cycle broker read timeouts (2026-11-19 09:30Z, 2026-11-20 09:29Z), each recovered in 6 s; co-core 0.50.2 on the index, version-only over this closure; refresh PR #214 open, CI green.
Activity
gregoryfoster commentedon Oct 26, 2026
#135 is now this thread's to coordinate (operator, 2026-10-26). The hand-off is #135 (comment); #135's body has been updated through
epic-edit.shand reads back clean.Merge #176: bump astral-sh/setup-uv from 10.3.2 to 10.3.3
gregoryfoster commentedon Nov 5, 2026
cannobserv v0.41.0: two lines for Replicator
v0.41.0 is tagged at cannobserv
f7285e1e. Replicator pins>=0.40.1,<0.41(main7c151486), so taking it is a floor move.BlobStorereporting twins (cannobserv#530).store_reporting/store_file_reportingreturnStored(uri, created), and the basestore/store_filenow returnself.…_reporting(...).uriby dispatch.SweepSafeLocalBlobStore(src/storage/local_store.py:33-39) overridesstore/store_fileonly, and it keeps working unchanged, sincesuper().storestill runs inside its retry. Before Replicator calls a twin, move the retry ontostore_reporting/store_file_reportingand drop the two plain-verb overrides. Keeping both nests the retry, so one write is attempted up to four times. A call to a twin on today's subclass skips the retry. The test-onlyThreadRecordingStore(tests/worker/test_storage_offloop.py), which overridesstorealone, wouldn't record a reporting call.FetchPolicyEmit.generationis required (cannobserv#579). This touches only the Emit side. Replicator decodesFetchPolicyState, which still takes a frame without one, andfetch_policy_supersedeskeeps its ratchet. Replicator has nothing to change.[-][Epic] Backlog burn-down: live stability, sibling waits, operational conformance (2026-10)[/-][+][Epic] Parking lot (burn-down done 2026-11-16): live stability, sibling waits, operational conformance[/+]#668 docs: coordinating-epics process-log entry, 2026-11-16 replicator