Repository navigation
docs(policy): supersede release-batched Vercel retention with active ordinary mode - #1031
Conversation
…ordinary mode The maintainer revoked the release-batched Preview retention record from #1020 and named the local Claude Code CLI session as the executing writer (GH #627). Add a dated Active mode section that supersedes the record without rewriting it, and point AGENTS.md and the deployment, merge and Cursor Cloud runbooks at it so no operative text still names Codex CLI as the destructive writer. The section does not reuse the #1020 bootstrap or declare another exception. Each deletion names one exact deployment ID revalidated against live data, a project name is never passed to vercel remove, and any 403, 429, 5xx, contradictory state or moved main stops the run. Tag and publish authority is not granted. The policy test pins the successor record and the pointer in every companion document.
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
🤖 CodeAnt AI — Review Status
|
Thanks for using CodeAnt! 🎉We're free for open-source projects. if you're enjoying it, help us grow by sharing. Share on X · |
Reviewer's GuideThis documentation-and-test change makes ordinary per-merge Vercel Preview retention the active maintainer-authorized mode, names the local Claude Code CLI as its writer, preserves the prior release-batched record as history, updates all operational pointers, and adds tests to prevent policy drift. Sequence diagram for active per-merge Vercel retentionsequenceDiagram
participant Claude as ClaudeCodeCLI
participant Vercel
participant GitHub
Claude->>Vercel: enumerate deployments and aliases
Claude->>Vercel: revalidate exact dpl_ ID before deletion
alt SAFE_TO_DELETE candidate
Claude->>Vercel: vercel remove exact dpl_ ID
Claude->>Vercel: re-enumerate to verify outcome
else no SAFE_TO_DELETE candidate
Claude-->>GitHub: record complete no-candidate pass
end
Claude-->>GitHub: record SHA, run IDs, counts, deleted IDs, verification
Flow diagram for retention policy supersessionflowchart LR
A[Release-batched retention record #1020] -->|superseded 2026-10-13| B[Ordinary per-merge retention]
B --> C[Local Claude Code CLI writer]
B --> D[Full inventory and live revalidation]
B --> E[Fail-closed deletion and alias rules]
F[AGENTS.md and operational runbooks] -->|point to Active mode| B
File-Level Changes
Possibly linked issues
Tips and commandsInteracting with Sourcery
Customizing Your ExperienceAccess your dashboard to:
Getting Help
|
🏁 CodeAnt Quality Gate ResultsCommit: ✅ Overall Status: PASSEDQuality Gate Details
|
|
Warning Review limit reachedEnable usage-based reviews in Billing to review now. Otherwise, wait until the next included review is available. Next included review available in 31 minutes. View limit details
📝 Walkthrough
Priority: ⬇️ Low Merge Risk: 🔵 Low · up to
|
CodeAnt Nitpicks1 code suggestion1. The two added policy tests move the stated total from 8543 to 8545, so this 8547+ metric overstates the test count by two.Logic error · |
CodeAnt PR Risk: Medium Risk
Assessed commit: |
Codecov Report✅ All modified and coverable lines are covered by tests. 📢 Thoughts on this report? Let us know! |
# Conflicts: # CHANGELOG.md # README.md
|
@CodeAnt-AI review |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 2bfc3805dd
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
…r implements Revalidation is per ID before its first write, not before every individual call; a branch alias is removed only when its branch is gone; a write is never retried, and only reads may retry once. Pins the corrected claims in the policy test.
|
@CodeAnt-AI review |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 915b4b298b
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
There was a problem hiding this comment.
Actionable comments posted: 2
ℹ️ Review info
⚙️ Run configuration
- Configuration used: Repository: qnbs/WorldScript-Studio/.coderabbit.yaml
- Review profile: CHILL
- Plan: Essentials
- Run ID:
f12df674-5f47-41e6-bd71-9b24a1a04363
📒 Files selected for processing (3)
README.mddocs/VERCEL-PREVIEW-RETENTION-POLICY.mdtests/unit/vercelRetentionTrainPolicy.test.ts
🚧 Files skipped from review as they are similar to previous changes (1)
- README.md
Included review availability: This review used your included allowance. 1 included review remains after this review. Your included PR review attempts over the past 7 days set your current allowance at 2 reviews per hour.
… the stop rule A failed read is retried once and stops the run on a second failure; a failed write is never retried and stops the run unless a re-read shows the removal landed, as the runner settles an alias or deployment write of unclear outcome.
|
@CodeAnt-AI review |
There was a problem hiding this comment.
Gates Passed
3 Quality Gates Passed
See analysis details in CodeScene
Quality Gate Profile: The Bare Minimum
Install CodeScene MCP: safeguard and uplift AI-generated code. Catch issues early with our IDE extension and CLI tool.
User description
Summary
Supersedes the release-batched Vercel Preview retention record from #1020 with a dated Active mode section, and points
AGENTS.md,docs/DEPLOYMENT.md,docs/PR-CI-MERGE-WORKFLOW.mdanddocs/CURSOR-CLOUD-AGENT.mdat it, so that no operative text still names Codex CLI as the destructive retention writer.Why
The maintainer revoked the release-batched mode and named the local Claude Code CLI session as the executing writer of ordinary per-merge retention (decision recorded in GH #627, comment 6134785538, and QNB-67). The #1020 record's own abort clause already treats a maintainer revocation as an abort into ordinary reconciliation, but the runbooks kept naming a Codex-only writer and an exceptional Codex handoff. The first ordinary pass under this mode is documented in GH #627 (comment 6135201825).
What changes
docs/VERCEL-PREVIEW-RETENTION-POLICY.md: new section "Active mode from 2026-10-13: ordinary per-merge retention" withACTIVE_MODE,ACTIVE_DESTRUCTIVE_WRITER,SUPERSEDES,SUPERSESSION_AUTHORITYand six writer predicates (exactdpl_IDs only, never a project name forvercel remove; complete inventory; live revalidation before each deletion or alias removal; branch-alias proof; fail-closed stops; a no-candidate pass is complete). The docs(policy): batch Vercel preview retention until release #1020 record stays unchanged as history.tests/unit/vercelRetentionTrainPolicy.test.ts: two cases pin the successor record and the pointer in every companion. README test counts follow (8543 → 8545).Not changed
This does not reuse the #1020 bootstrap, does not declare another exception, and does not waive any post-merge requirement. It grants no tag, publish, Gate 7 or production storage-authority authority. The bounded dependency/toolchain train is unchanged.
Summary by cubic
Supersedes the release-batched Vercel Preview retention record with a dated Active mode section that names the local Claude Code CLI session as the destructive writer, per the maintainer decision in GH #627.
## Active mode from 2026-10-13section todocs/VERCEL-PREVIEW-RETENTION-POLICY.md, keeping the docs(policy): batch Vercel preview retention until release #1020 record unchanged as history.dpl_deployment ID, never a project name; each ID is revalidated against live data before its first write, a branch alias is removed only when its branch is gone, a failed read is retried once and a second failure stops the run, and a failed write is never retried and stops the run unless a re-read shows the removal landed. Incomplete inventory or a movedmainstops the run.AGENTS.md,docs/DEPLOYMENT.md,docs/PR-CI-MERGE-WORKFLOW.md, anddocs/CURSOR-CLOUD-AGENT.mdat the new section so no operative text still names Codex CLI as the retention writer; Cursor Cloud still does not delete deployments.CHANGELOG.mdentry, plus two policy tests pinning the successor record and the pointer in every companion; README test counts follow (8544 → 8547).Written for commit 0bf6ae7. Summary will update on new commits.
Summary by Sourcery
Adopt ordinary per-merge Vercel Preview retention as the active policy and align its documentation and tests with the maintainer-approved execution model.
New Features:
Enhancements:
Documentation:
Tests:
Summary by CodeRabbit
CodeAnt-AI Description
Document ordinary per-merge Vercel Preview cleanup as the active mode
What Changed
main, carried out by the local Claude Code CLI operator; the old record remains as history.Impact
✅ Per-merge Preview cleanup is the active procedure✅ Fewer risks from broad or stale deployment deletions✅ Cleanup stops when provider state is unclear💡 Usage Guide
Checking Your Pull Request
Every time you make a pull request, our system automatically looks through it. We check for security issues, mistakes in how you're setting up your infrastructure, and common code problems. We do this to make sure your changes are solid and won't cause any trouble later.
Talking to CodeAnt AI
Got a question or need a hand with something in your pull request? You can easily get in touch with CodeAnt AI right here. Just type the following in a comment on your pull request, and replace "Your question here" with whatever you want to ask:
This lets you have a chat with CodeAnt AI about your pull request, making it easier to understand and improve your code.
Example
Preserve Org Learnings with CodeAnt
You can record team preferences so CodeAnt AI applies them in future reviews. Reply directly to the specific CodeAnt AI suggestion (in the same thread) and replace "Your feedback here" with your input:
This helps CodeAnt AI learn and adapt to your team's coding style and standards.
Example
Retrigger review
Ask CodeAnt AI to review the PR again, by typing:
Check Your Repository Health
To analyze the health of your code repository, visit our dashboard at https://app.codeant.ai. This tool helps you identify potential issues and areas for improvement in your codebase, ensuring your repository maintains high standards of code health.