A final review agent for high-stakes work, including software builds, AI and machine-learning tasks, automation, operations, executive advice, and personal finance. It checks whether the result has enough fresh evidence before approving it.

Use it to audit high-risk outputs and changes to installation scripts, hooks, GitHub workflows, or secrets. It records its verdict and evidence in an audit log.

It reduces the chance of treating incomplete or weakly verified work as finished. When evidence is missing, it sends the work back with specific fixes.

auditor — Agent for Claude Code by Momo2323-ui/claude-orchestra

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Clone the repo

git clone --depth 1 https://github.com/Momo2323-ui/claude-orchestra

Made for: Claude Code.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

README.md

[![agentmods](https://agentmods.dev/badge/agents/momo2323-ui/claude-orchestra/auditor/github.svg)](https://agentmods.dev/agents/momo2323-ui/claude-orchestra/auditor)

Your own site

<a href="https://agentmods.dev/agents/momo2323-ui/claude-orchestra/auditor"><img src="https://agentmods.dev/badge/agents/momo2323-ui/claude-orchestra/auditor/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

Your own site · 80×15

<a href="https://agentmods.dev/agents/momo2323-ui/claude-orchestra/auditor"><img src="https://agentmods.dev/badge/agents/momo2323-ui/claude-orchestra/auditor.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 113 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.

When invoked 3,481 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 1 finding. A grade says what 26 rules found in the file — not that it is safe.

Origin original No closer match found in the catalogue.

Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

Model	Per session	Once invoked
Fable 5.1	$0.00113	$0.03481
Opus 5.5	$0.00045	$0.01392
Sonnet 5.5	$0.00023	$0.00696
Haiku 4.5	$0.00011	$0.00348

Measured 2mo ago against content hash 982fc4e2af2f, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

auditor scanned grade A with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2mo ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Reads agent configuration directorieslowAgent snooping

.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.

- ANY change touching: `.github/workflows/`, `install.sh`, `~/.claude/hooks/`, `~/.claude/settings.json`, anything secret-handling

Downgraded: this mod is about security review, or the phrase is quoted, so it is likely naming the pattern rather than instructing it.

agents/auditor.md · 285 lines

How it starts

The opening of the file, as written. The whole thing — 285 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are the conductor of AUDIT — orchestra #21. You are the last gate before any high-stakes orchestra output reaches the user. Your job is to default to NEEDS WORK, demand fresh evidence, and either flip to READY or bounce the work back to the source orchestra with a specific fix list. You do NOT implement fixes. You find what's missing, you route, you re-audit.

You answer to the user. You protect the user from work that sounds done but isn't.

WRITE BOUNDARY (hard rule): your Write tool exists for exactly ONE path — ~/.claude/docs/learnings/audit-log.md (append your verdict + evidence). Never write, edit, or create any other file. You are read-and-inspect everywhere else; fixes belong to the source orchestra, not to you. If a fix is needed, put it in the bounce-back list.

YOUR PLACE IN THE FLOW

Source orchestra (e.g. BUILD) → produces output + evidence
   ↓
AUDIT (you) → defaults to NEEDS WORK
   ├─ Evidence sufficient? Verdict = READY → return to user with provenance
   └─ Evidence insufficient? Verdict = NEEDS WORK
        → bounce back to source orchestra with specific fixes
        → max 3 attempts → escalate via 🤚 ASK_HUMAN

ENTRY DISCIPLINE — Iron Law applies to YOU first. On entry, your first action is Skill(skill: "hallucination-guard"). No exceptions. No "if it hasn't fired yet" — invoke it every audit, before reading any source-orchestra output, before forming any verdict, before opining on anything. You are running the Iron Law on your own work, not just on the source orchestra's. Exit the same way — invoke it again before emitting the verdict.

ACTIVATION — WHEN AUDIT FIRES

High-stakes orchestras (audit fires by default)

① BUILD — any code change
⑮ AI/ML DEVELOPMENT — model code, RAG pipelines, agent system changes
⑭ AUTOMATION & OPS — hooks, n8n flows, runbooks, CI
⑲ PERSONAL FINANCE & INVESTING — any analysis presented as fact
⑳ EXECUTIVE ADVISORY — any decision-framing output
ANY change touching: .github/workflows/, install.sh, ~/.claude/hooks/, ~/.claude/settings.json, anything secret-handling

Read the full file on GitHub · 285 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

2mo ago First seen · 285 lines · 113 tokens per session scan A 982fc4e2af2f

Subscribe to this mod's changes

auditor is an agent published in the GitHub repository Momo2323-ui/claude-orchestra (9 stars, last pushed 3mo ago), licensed MIT. It adds 113 tokens to every session and 3,481 once invoked, about $0.0005 per session on Opus 5.5. A static security scan graded it A with 1 finding (reads agent configuration directories). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Claude Code Code review agent-skills ai-agents anthropic claude developer-tools

auditor next to 6 agents from other repositories
Mod	Type	Stars	Last push	Licence	Scan	Tokens	Works with
auditorMomo2323-ui/claude-orchestra	Agent	9	3mo ago	MIT	A	113	Claude Code
security-revieweraffaan-m/ECC	Agent	299k	yesterday	MIT	A	52	Claude Code, Kiro
cpp-revieweraffaan-m/ECC	Agent	299k	yesterday	MIT	A	41	Claude Code, Kiro
component-improverdavila7/claude-code-templates	Agent	33k	today	MIT	A	39	Claude Code
thermo-nuclear-code-quality-review-subagentmichael-denyer/pstack-claude	Agent	4.1k	yesterday	MIT	A	68	Claude Code
predictive-analystsoftspark/ai-toolkit	Agent	179	yesterday	Apache-2.0	A	35	Claude Code
atomic-auditordamusix/atomic-claude	Agent	116	5d ago	MIT	A	169	Claude Code

auditor