Lints SKILL.md files and scripts for credentials, broken references, and convention violations.
npx skills add https://github.com/cpanse/skills --skill skill-audit-cpanse
Please help me install this Agent Skill. Skill: skill-audit Source: https://github.com/cpanse/skills/tree/main/meta-skills/skills/skill-audit Command: npx skills add https://github.com/cpanse/skills --skill skill-audit-cpanse
--threshold P1 to fail the pipeline if the contributor committed a MySQL password or a /home/<username>/ path. ## Quick Start Run the audit script on the marketplace to list all P0 and P1 findings before committing my skill changes.High-intent search queries and answers about installing and using this skill.
Run scripts/audit.py from the marketplace root; the no_plaintext_credentials rule detects mysql -p passwords, hardcoded password assignments, API keys, and Bearer tokens. Use --threshold P0 in a pre-commit hook to block commits containing secrets.
Add a test stage job using a python:3.11-slim image that runs audit.py with --threshold P1 and allow_failure: false. The script exits with code 1 when findings meet the threshold, failing the pipeline on P0 and P1 issues.
No, audit.py uses only the Python standard library with no external dependencies and no network access. It runs on any Python 3 installation, making it safe for CI environments and offline clusters.
Regex rules cannot catch semantic issues like citation accuracy, scientific correctness, logic errors, or general API drift beyond hardcoded cases. Those require the companion multi-LLM review approach; the linter is the fast mechanical first pass.
Rules filter common placeholder shapes like fgcz-r-NNN, <username>, and ${VAR} to keep false positives low. When a true false positive appears, the fix is refining the rule regex in audit.py rather than adding per-file ignores.
SethyPagnaStandardizes creation and validation of in-repo JARVIS SKILL.md files.
tizeeDetect layer boundary violations and dependency-direction leaks with P0-P3 severities.
crinisReviews CSS code for redundancy, duplication, and semantic consistency with linters.
OWASPDetect hardcoded secrets in source code and configuration files.