auditor
Before you commit, publish, or trust something — a change, a claim, a plan, a "this is done."
You need to understand something before acting — explore a codebase, gather sources, map how a system works.
How it fires
How this agent gets triggered: by you, by Claude, or both.
Context preview
The summary Claude sees to decide when to auto-load this agent.
You need to understand something before acting — explore a codebase, gather sources, map how a system works.
You need to understand something before acting — explore a codebase, gather sources, map how a system works.
A findings map: `claim → evidence (path/link) → confidence`. Anything surprising is flagged for you to spot-check before you rely on it.
Research in its own context keeps your main thread clean, and a read-only mandate prevents "while I was in there I also changed…" drift. (Ties to the rule: *agent output is input, never the conclusion.*)
Stop fixing the same AI mistake twice. A Lean quality system for AI-assisted work — operators turn recurring defects into countermeasures. Built by a manufacturing operator, for people who actually ship.
Before you commit, publish, or trust something — a change, a claim, a plan, a "this is done."
You're scattered, stacked with options, or unsure what to do next.
Your memory folder is growing, the index is getting long, or facts feel duplicated or stale.