npx agentspec-cli install f872e47b-4d24-4d97-b7eb-549696b5eddc --framework ClaudeCode→ .claude/skills/ap/SKILL.md
# Accounts Payable (read-only) Package `ap_engine` in `accounts-payable-automation/`. ## Commands (run from `accounts-payable-automation/`) ```bash python run.py # generate the fictional corpus, run all 30 controls, write both artifacts python -m ap_engine ./samples # analyze an existing folder read-only python -m pytest -q # 12,223 tests incl. the planted-defect suite ``` Exit codes: `0` PASS · `1` REVIEW · `2` FAIL · `3` usage/IO. The bundled corpus carries one planted defect per control, so it exits `2` by design. ## The one rule This engine NEVER writes — source artifacts are hashed before and after a run and asserted byte-identical. It has no connection to a ledger, payment provider or approval system, so nothing it produces can post or pay. Do not bolt a remediation loop onto it; like `/validate` it is the *sensor* other loops use. If asked to "fix" what it finds, route the fix to the source system, then re-run. ## What to watch for - **Absent evidence is not a passing control.** `set_complete` runs first for exactly this reason: a set missing an artifact type would otherwise report clean because the controls reading it returned nothing. - **A benign notice is not a failure.** A "job-cost entries not created" style line is normal for a ledger-only invoice; only a blocking marker (batch contention, aborted update) is a FAIL. Conflating them turns the control into noise. - **PASS is not sign-off.** It means the implemented mechanical controls found nothing. Release, approval and posting stay with a person. - Money is integer cents with exact `==` and no tolerance band anywhere. A non-integer amount is rejected, never coerced.
Built on the Agent Skills standard
minimal-run-and-audit
Trusted-lane execution and reporting skill for README-first AI repo reproduction. Use when the task is specifically to capture or normalize evidence from the selected smoke test or documented inference or evaluation command and write standardized `repro_outputs/` files, including patch notes when repository files changed. Do not use for training execution, initial repo intake, generic environment setup, paper lookup, target selection, or end-to-end orchestration by itself.
by AgentHub Crawler
=== USER INSTRUCTIONS ===
The license management system implements a hardware-locked commercial licensing solution with tiered access control and multi-stage validation.
by AgentHub Crawler
audit
Run technical quality checks across accessibility, performance, theming, responsive design, and anti-patterns. Generates a scored report with P0-P3 severity ratings and actionable plan. Use when the user wants an accessibility check, performance audit, or technical quality review.
Sign in to leave a comment.
by AgentHub Crawler