Installs into .claude/skills of the current project.
Sponsors
SSkills DirectorySkills Directory
Know which skills are safe — weekly.
SSkills DirectorySkills Directory
Is that skill safe to install?
Are you the author of Audit Protocol?
Add the live security badge to your README. It updates with every re-scan.
[](https://www.skillsdirectory.com/skills/sarj-ai-audit-protocol-fdfd4abb)
---
name: audit-protocol
description: Run a requested domain audit using Sarj's shared command sources and verified findings.
---
Audits are read-only unless implementation is requested. Load the selected
`../../commands/` file relative to this skill; Claude: `/sarj-skills:<command>`.
Identify owning repositories, stack and boundaries; search named source paths
with bounded output. Exclude unrelated task outputs, caches, duplicate worktrees
and generated/vendor/fixture data. Run native checks before judging
uncovered behavior; inspect framework guards.
Require completed exit statuses; interrupted checks are inconclusive.
Keep reports ignored and private; redact secrets/identifiers.
Report severity, location, evidence, impact, smallest fix, confidence and owning
tool/rule or `judgment-only`; report clean scopes too.
For interruption, save objective, worktree, base/head/tree, commands/statuses,
reports/PRs, owned processes and next action. Reconcile revisions before reuse;
reference private history instead of copying it.
