audit-discovery
Skill1 starsChanged 35 days agoMITView on GitHub
Use when findings should be recorded as structured audit tickets. Cold-start friendly: initialize the audit workflow, inspect evidence, then create DRAFT/OPEN tickets. Do not resolve or verify tickets.
In this repository27 other files it gives its agents.
- architectSkill
- audit-initSkill
- audit-nextSkill
- audit-resolutionSkill
- audit-statusSkill
- audit-triageSkill
- audit-verificationSkill
- code-to-prdSkill
- codex-ubuntu-serverSkill
- copilot-review-customizerSkill
- create-pluginSkill
- debugSkill
- deep-reviewSkill
- design-promptsSkill
- diagnose-sessionSkill
- feature-scatteringSkill
- human-execution-runbookSkill
- humanizerSkill
- init-repo-instructionsSkill
- inspect-pluginSkill
- install-agent-pluginsSkill
- local-wikiSkill
- recoverSkill
- review-feedbackSkill
- review-repo-instructionsSkill
- uncompromising-handoverSkill
- update-pluginSkill
The file
We split this skill into the rules it gives an agent
9 rules. About 0 are common, 4 are distinctive, and 5 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.
5
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 9
Read the runtime invocation contract first.
- Distinctive: line 28
Use the Python CLI by default (or equivalent tools if MCP was configured manually):
- Very specific: line 36
If audit/ does not exist, that is normal.
- Very specific: line 56
create DRAFT or OPEN tickets in Ticketize or Audit sweep.
- Distinctive: line 58
add concrete evidence, acceptance criteria, module/line references, and suggested verification.
- Distinctive: line 62
do not implement fixes;
- Very specific: line 63
do not set READY_FOR_VERIFICATION.
- Very specific: line 64
do not set PASS, PARTIAL, FAIL, REGRESS, WONTFIX, or INVALID except through the appropriate role.
- Distinctive: line 65
do not invent tickets from vague suspicions.
≈ “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
Audit Discovery
Read the runtime invocation contract first. Use the bundled Python CLI after resolving AUDIT_PLUGIN_ROOT and AUDIT_PROJECT_DIR as described there. A separately configured audit_* MCP server is an optional equivalent with an explicit absolute root.
Purpose: convert concrete findings into audit tickets. The CLI/MCP runtime is the source of truth; Markdown files are the human-readable record.
Modes
| Mode | Write audit files? | Behavior |
| Investigate | No | Read code/logs and explain what is known. |
| Ticketize | Yes | Create tickets only for evidence-backed findings. |
| Audit sweep | Yes | Systematically inspect, then ticketize concrete findings. |
Choose the mode from the user's request before invoking the runtime. A request
to investigate or explain an issue uses Investigate: do not initialize audit/,
create/open tickets, or write audit files. Report the findings and uncertainties.
Ticketize and Audit sweep apply when the user requests audit records.
First move for Ticketize or Audit sweep
Use the Python CLI by default (or equivalent tools if MCP was configured manually):
python3 "${AUDIT_PLUGIN_ROOT}/scripts/audit.py" --root "${AUDIT_PROJECT_DIR}" init
python3 "${AUDIT_PLUGIN_ROOT}/scripts/audit.py" --root "${AUDIT_PROJECT_DIR}" doctor
python3 "${AUDIT_PLUGIN_ROOT}/scripts/audit.py" --root "${AUDIT_PROJECT_DIR}" summaryIf audit/ does not exist, that is normal. Initialize it through the runtime;
do not create the structure by hand. These startup steps do not apply to Investigate.
Create tickets
Incomplete finding:
python3 "${AUDIT_PLUGIN_ROOT}/scripts/audit.py" --root "${AUDIT_PROJECT_DIR}" create BUG --title "parse failure on null input" --severity high --module src/parser.pyEvidence-ready ticket:
python3 "${AUDIT_PLUGIN_ROOT}/scripts/audit.py" --root "${AUDIT_PROJECT_DIR}" create BUG --title "parse failure on null input" --severity high --module src/parser.py --description "Parser crashes when payload.user is null." --evidence "src/parser.py:42 dereferences payload['user']['id'] without a guard" --acceptance-criterion "Null user payload returns a typed validation error" --suggested-verification "Reproduce null-user payload and verify typed validation error" --open --jsonAllowed actions
- initialize the workflow in Ticketize or Audit sweep;
- create
DRAFTorOPENtickets in Ticketize or Audit sweep; - promote evidence-ready
DRAFTtickets toOPENusingpython3 "${AUDIT_PLUGIN_ROOT}/scripts/audit.py" --root "${AUDIT_PROJECT_DIR}" open ... --as audit-discovery; - add concrete evidence, acceptance criteria, module/line references, and suggested verification.
Forbidden actions
- do not implement fixes;
- do not set
READY_FOR_VERIFICATION; - do not set
PASS,PARTIAL,FAIL,REGRESS,WONTFIX, orINVALIDexcept through the appropriate role; - do not invent tickets from vague suspicions.
Stop condition
In Investigate, stop after reporting evidence and uncertainties. In Ticketize or
Audit sweep, stop after tickets are created/opened and hand off to audit-triage
or audit-resolution.
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.

