I build small command-line tools that solve one real problem each, in pure Python with no dependencies: nothing to install, nothing to audit, nothing that phones home. Every tool is a single repo with tests, a README, CI and an MIT licence, and every one is defensive or protective: they check, warn and report. None of them attacks anything.
pip install git+https://github.com/TPOALOS/<tool> # any tool belowBrowse everything in one place: tpoalos.github.io
| Tool | What it does | |
|---|---|---|
| 🛡️ | hostguard | Read-only hardening audit for Linux and macOS: 76 checks, a score, an HTML report and drift detection |
| 🎯 | detectlite | Detection engineering: a Sigma-compatible rule engine for local logs with 42 tested rules mapped to MITRE ATT&CK |
| 🔎 | portlens | Defensive attack-surface scanner (nmap-style) with a scope guard and rate caps: it refuses to scan what you do not own |
| 📦 | pcaplens | pcap/pcapng analyzer: TCP reassembly, scan, beacon and DNS-tunnel detectors, CTF stream tools |
| 🏁 | ctfhost | Host your own CTF: scrypt logins, CSRF protection, dynamic scoring, first blood, a scoreboard |
| 💰 | pocketledger | Private, local personal finance: bank CSV import, budgets, subscription detection, net worth, an HTML report |
Eight finance tools that fit together as one workflow for testing a trading idea honestly. All are offline, dependency-free and read your CSVs:
ohlcvlint -> backtestlint -> edgelab -> riskmetrics -> costmap -> slipcheck -> ruinsim
is the data is the code is the edge is the Sharpe where does what did how big can
sane? leak-free? real? real? it die? costs be? I trade?
strategy-graveyard: what would pure luck have done on the same numbers?
- ohlcvlint: check the price data for bad bars, gaps and unadjusted splits before anything else.
- backtestlint: lint the backtest code for look-ahead and leakage before you trust any number.
- edgelab: log every variant you try, then check the holdout, the trial count and the cost.
- riskmetrics: confidence intervals and a deflated Sharpe for the winner.
- costmap: net Sharpe across costs and volatility regimes, to see where the edge dies.
- slipcheck: measure what your fills really cost and compare it with the cost you assumed.
- ruinsim: bootstrap your trades to size positions against drawdown and ruin.
- strategy-graveyard: run the classic rules on no-edge random markets to see how good luck can look.
| Tool | The question it answers |
|---|---|
| cmdguard | Will this command my AI agent wants to run wreck my machine? (allow, ask or deny; works as a Claude Code hook) |
| mcpaudit | Is my MCP server configuration safe? (secrets, unpinned downloads, over-broad access) |
| injectscan | Is there a hidden instruction in this page or file? (invisible Unicode, hidden HTML, decoded for you) |
| ctxpack | How do I give an AI my whole project without leaking my secrets? |
| tokcount | How many tokens is this, will it fit, and what will it cost? |
| Tool | The question it answers |
|---|---|
| backupcheck | Is my backup really there, fresh, not empty, and readable to the end? |
| jobwatch | Did my cron job silently stop running? |
| diskwatch | Will this disk fill up, and in how many days? |
| expirywatch | What expires soon: passport, domain, certificate, insurance? |
| certwatch | When do my TLS certificates expire? |
| cronnext | What does this crontab really do, and when does it run next? |
| dupefind | Which files are duplicates, and how much space do they waste? |
| Tool | The question it answers |
|---|---|
| exifstrip | What does this photo give away (GPS, camera, name), and can I remove it? |
| secretscan · githook-guard | Did I just commit a secret? Block it before it leaves my machine |
| gitidentity-audit | Which of my commits expose a personal e-mail address? |
| logscrub | Redact secrets, e-mails and IPs from a log before I share it |
| shellhist-scrub | Which passwords did I type into my shell history? |
| envdrift | Does my .env match .env.example, and did I leave a placeholder secret? |
| pwlint | How weak is this password, really? (it is never printed or sent) |
| pwned-local · pwhash-audit | Is this password in a breach list (offline)? Does my shadow or htpasswd file hold weak hashes? |
| Tool | The question it answers |
|---|---|
| listenwatch | What is listening on this machine, and who can reach it? |
| sshcfg-audit · sshkey-audit | Is my SSH server, client and key set-up safe? |
| authlog-summary | Who is hammering my SSH, and did one of them get in? |
| fimcheck | Did any file change, or its permissions or owner? |
| permaudit | Which files are world-writable, SUID or have loose permissions? |
| headercheck | How good are my site's security headers? |
| jwtlint | What is wrong with this JWT? |
| mailtriage · urlcheck | Is this e-mail or link a phishing attempt? (offline, never opened) |
| pathaudit · symlinkcheck · sudoers-audit | Can someone hijack my PATH, my symlinks or my sudo rules? |
| elfcheck · iocgrep | Is this binary hardened (PIE, NX, RELRO)? Do these logs contain known indicators of compromise? |
| Tool | The question it answers |
|---|---|
| nginxcheck · cookiecheck | Is my nginx config or my Set-Cookie header weak? |
| mailauth-lint · dmarc-report | Are my SPF, DKIM and DMARC records right, and who is failing them? |
| zipcheck · csvguard · trojansrc | Is this archive, spreadsheet or source file hostile? |
| Tool | The question it answers |
|---|---|
| repolint | Is this repository ready to share? |
| gha-audit | Is my GitHub Actions workflow safe? |
| dockerfile-audit | Is my Dockerfile safe? |
| reqpin | Are my pip requirements pinned, hashed and free of typosquats? |
| npmscan · signcheck | Does my npm project run install scripts or pull odd dependencies? Are my commits signed? |
| Tool | The question it answers |
|---|---|
| planguard · iamlint | What will this Terraform plan destroy or expose? Is this IAM policy too broad? |
| apispec-lint · zonelint | Is my OpenAPI spec or DNS zone file weak? |
| Tool | The question it answers |
|---|---|
| backtestlint | Does my backtest peek into the future? (look-ahead and leakage lint) |
| riskmetrics | Is this Sharpe ratio real, or the luckiest of many tries? (deflated Sharpe, drawdowns) |
| edgelab | Is this trading edge real? A trial ledger counts every variant, then checks the holdout and the cost |
| costmap | Where does my strategy die? Net Sharpe by transaction cost and volatility regime, with no look-ahead labels |
| strategy-graveyard | What does luck look like? Classic rules on no-edge random markets, and does my price history beat them? |
| ohlcvlint | Is my price data sane? Bad bars, gaps, unadjusted splits, spikes and stale prices, found before they fake a backtest |
| slipcheck | Did my backtest assume the cost I actually paid? Real slippage and fees from my fills, with an interval |
| ruinsim | How big can I trade before the drawdown kills me? Bootstrapped drawdown and ruin risk by position size |
- Zero dependencies. Python's standard library only, so there is nothing to install and nothing to audit.
- Read-only by default. The tools report; when one can change something (
--strip,--scrub,--append) it previews first, keeps a backup or writes a copy, and never deletes on its own. - Honest about limits. These are checks, not oracles: a clean result means "nothing found", never "proven safe", and the tools say so.
- Tested like it matters. More than 2,000 tests across the repos, including hostile inputs (paths with
$(...), newlines in file names, symlinks that point outside the folder, truncated archives). The maintained tools run CI on Linux and macOS with actions pinned to commit SHAs. - Private by construction. No network access, no telemetry, no accounts. Tools that must look at a network (
certwatch,headercheck,portlens) only ever talk to hosts you name.
Issues and pull requests are welcome. If a tool saved you from a bad day, a star helps others find it.
