audit

/audit — Spawn the Auditor Agent

The /audit skill is a thin wrapper around agents/auditor.md. The auditor agent file holds the canonical methodology (per-story flow, sprint-close branch sweep, PRD review mode, three canonical rules, common miss patterns, audit report template, sign-off protocol, wrap protocol). This skill exists for ad-hoc invocations outside the Ralph loop.

When this skill fires

• User explicitly says /audit, "audit the sprint", "review the coding agent's work", "set up the audit agent", or any mention of verifying stories against acceptance criteria.
• A teammate audit is needed but the lead isn't in a Ralph loop (one-off audit of an arbitrary commit, post-merge re-audit, etc.).

Inside an active Ralph sprint with Audit.mode = per-story, the lead spawns the auditor automatically after each feat: commit per /start's Audit Gate — you don't need to invoke this skill manually.

What this skill does

Team name (<team>): resolve once per session — team_name from RALPH.md's ## Audit section if set, otherwise ralph-<project-folder> (the kebab-cased basename of the repo root, e.g. ralph-closetize-website). Never use the bare name ralph: teams are machine-global (~/.claude/teams/), shared across every terminal and directory, so a fixed literal name cross-wires concurrent Ralph projects — a lead in one repo can wake an idle session in another repo and send it work.

Spawn the auditor agent and hand off to it.

Agent({
  subagent_type: "auditor",
  team_name: "<team>",
  name: "auditor",
  description: "Ralph sprint auditor",
  prompt: "Run setup per agents/auditor.md, then audit <STORY-ID> at commit <SHA>. Report verdict via SendMessage to team-lead."
})

If no team exists yet, create one first via TeamCreate({ team_name: "<team>", agent_type: "team-lead", description: "Ralph sprint coordination" }).

For ad-hoc use, ask the user what they want audited:

• A specific story ID + commit SHA → per-story flow
• "The branch" → sprint-close branch sweep
• A PRD draft path → PRD review mode

After spawning, the auditor handles the rest — it reads RALPH.md, the PRD, and PATTERNS.md, then runs whichever flow matches the request. Reports are written to the directory configured in RALPH.md Audit.reports_dir (default scripts/ralph/audit-reports).

Where the methodology actually lives

agents/auditor.md — single source of truth. This skill must NOT duplicate that content; if you find yourself wanting to add methodology details here, add them to the agent file instead.

Cross-references

• Agent file: agents/auditor.md
• Audit reports: <Audit.reports_dir>/R-*.md, BRANCH-AUDIT-*.md, PRD-AUDIT-*.md
• Auditor cross-session memory: <Audit.reports_dir>/PATTERNS.md
• Auditor wrap state: <Audit.reports_dir>/audit-progress.txt
• Sprint loop entry point: skills/start/SKILL.md (Audit Gate section)