coord-audit
Skill0 starsChanged 47 days agoApache-2.0View on GitHub
Run the CoordWright hardening audit — read-only check of codex sandbox config, auth file perms, worker.sh integrity, execpolicy allow-list, Claude Code hook installation, and per-project .gitignore + .env perms. Use when the user says "/coord-audit", "audit hardening", "check security posture", or wants a one-shot drift check.
- Reads credentials
In this repository20 other files it gives its agents.
- coord-archive-indexSkill
- coord-assignSkill
- coord-checkSkill
- coord-dirtySkill
- coord-discardSkill
- coord-helpSkill
- coord-initSkill
- coord-loop-doctorSkill
- coord-promoteSkill
- coord-refineSkill
- coord-requeueSkill
- coord-reviewSkill
- coord-shape-reviewSkill
- coord-shapeSkill
- coord-splitSkill
- coord-statusSkill
- coord-tokensSkill
- coord-unblockerSkill
- merge-to-mainSkill
- rebase-to-mainSkill
The file
We split this skill into the rules it gives an agent
8 rules. About 0 are common, 4 are distinctive, and 4 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈4
Distinctive · estimate
This author's own choices. Worth a look.
4
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 17
No flag: run the hardening audit only (fast.
- Distinctive: line 32
If the hardening audit exits non-zero, lead the response with the FAIL lines, then show the full output.
- Distinctive: line 33
If the secrets sweep is run, present its output AFTER the hardening audit so a FAIL is not buried under advisory noise.
- Very specific: line 36
Only if launchctl list | grep com.coord.audit.daily shows a manually installed nightly job, point the user at its log …
- Very specific: line 37
If any FAIL is in scope, point at the relevant project docs or memory for context (e.g.
- Distinctive: line 41
Never mutate any file from this skill.
- Very specific: line 42
If output is long, show the head and tail with a [...
- Very specific: line 44
If audit-hardening.sh is missing at the expected path, print that the CoordWright install needs to be re-run (bash …
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Read-only inspection of the operational hardening landed across the CoordWright checkout, ~/.codex/, ~/.claude/, and each registered project. Does not modify anything. This skill runs the check on demand. install.sh does not install a scheduled audit; a machine may have an optional, manually installed LaunchAgent (for example com.coord.audit.daily logging to ~/Library/Logs/coord-audit.log), but do not assume one exists.
Source of truth:
- Script:
${COORD_TOOLS:-$HOME/Projects/coord-wright}/bin/audit-hardening.sh - Companion sweep across projects (advisory, not gate):
${COORD_TOOLS:-$HOME/Projects/coord-wright}/bin/audit-projects-secrets.sh
Syntax
/coord-audit [--secrets]
- No flag: run the hardening audit only (fast; exits 0 if all checks pass, 1 if any fail).
--secrets: also run the project-wide secrets sweep (advisory; always exits 0; flags world-readable .env files, credential-pattern files, and tracked files matching secret-keyword regex).
Workflow
- Run the hardening audit.
bash "${COORD_TOOLS:-$HOME/Projects/coord-wright}/bin/audit-hardening.sh"- Capture exit code. The script prints PASS/FAIL/SKIP per check and a
N passed, M failedsummary line.
- If
--secretswas passed, also run the sweep.bash "${COORD_TOOLS:-$HOME/Projects/coord-wright}/bin/audit-projects-secrets.sh"- This is advisory — exits 0 even when findings exist. The user should triage each flagged file.
- Present output.
- Stream each script's output verbatim. Do not summarise individual checks; the script's own table is the answer.
- If the hardening audit exits non-zero, lead the response with the FAIL lines, then show the full output.
- If the secrets sweep is run, present its output AFTER the hardening audit so a FAIL is not buried under advisory noise.
- Cross-reference for context.
- Only if
launchctl list | grep com.coord.audit.dailyshows a manually installed nightly job, point the user at its log (~/Library/Logs/coord-audit.log) for drift between manual runs. Otherwise say nothing about scheduled audits;install.shdoes not install one. - If any FAIL is in scope, point at the relevant project docs or memory for context (e.g. the project's
AGENTS.md, or a relevant note under~/.claude/projects/<project>/memory/).
- Only if
Rules
- Never mutate any file from this skill. For chmod fixes or sandbox changes, the user makes the call explicitly.
- If output is long, show the head and tail with a
[... N lines truncated ...]marker and always keep FAIL/WARN lines. - Report results plainly; do not fabricate check names. The script defines what is checked.
- If
audit-hardening.shis missing at the expected path, print that the CoordWright install needs to be re-run (bash "${COORD_TOOLS:-$HOME/Projects/coord-wright}/install.sh") and stop.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

