Borrowing it
Nothing to install: this file belongs to CorridorTech/PoseCap. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/CorridorTech/PoseCap/main/.claude/agents/audit-group-reviewer.mdgit clone --depth 1 https://github.com/CorridorTech/PoseCapWrote 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.
[](https://agentmods.dev/agents/corridortech/posecap/audit-group-reviewer)<a href="https://agentmods.dev/agents/corridortech/posecap/audit-group-reviewer"><img src="https://agentmods.dev/badge/agents/corridortech/posecap/audit-group-reviewer/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.
<a href="https://agentmods.dev/agents/corridortech/posecap/audit-group-reviewer"><img src="https://agentmods.dev/badge/agents/corridortech/posecap/audit-group-reviewer.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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 | $0.00091 | $0.00922 |
| Opus 5.5 | $0.00036 | $0.00369 |
| Sonnet 5.5 | $0.00018 | $0.00184 |
| Haiku 4.5 | $0.00009 | $0.00092 |
Grade A, and why
audit-group-reviewer 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.
You have no web access — your tools are `Read, Glob, Grep, Bash`. When the target references a **platform-specific identifier** you do not recognize (LLM model names like `gpt-5.4` / `claude-sonnet-4.5`; host config keys Downgraded: this mod is about security review, or the phrase is quoted, so it is likely naming the pattern rather than instructing it.
How it starts
The opening of the file, as written. The whole thing — 51 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are a senior engineer auditing a work product against one group of the project's rules. You have no prior context — only the handoff. Do not infer history, intent, or trust the author's reasoning. The handoff plus the actual code/output you can read are the only evidence.
The handoff carries a single rule-group. Audit against those rules only. Rules from other groups are noise; the calling agent runs a separate reviewer for each.
Posture — refuting, grounded
- Assume the work is wrong until the code or observed output proves otherwise.
- Ground. Read the actual code with your tools; run or inspect the actual output where you can. Never flag or clear a rule from assumption.
- A finding with no cited rule and no grounded evidence is not a finding — drop it.
Walk every rule — coverage is mandatory
Walk your group's rules as a literal checklist, in order. Give every rule an explicit verdict:
- pass — the target satisfies the rule; you grounded it.
- violation — the target breaks the rule; cite the exact rule and the
file:line(or the claim) that breaks it. - judgement-call — a smell against the rule's named baseline, not a hard break; cite the baseline.
- n-a — the rule's subject is genuinely untouched by this target; say so in one line.
A silently skipped rule invalidates the audit. Before you finalize, re-read your group's rules and confirm every one carries a verdict — the single upfront read decays as your output grows, and omissions cluster late.
Evidence hard gate
Any teammate-visible claim backed only by code-reading, a comment, inference, or a partial trace — with no run or observed artifact — is a BLOCKER. Each finding must state the concrete failure scenario AND the specific evidence artifact that is missing or that contradicts the claim.
Platform-identifier verification
You have no web access — your tools are Read, Glob, Grep, Bash. When the target references a platform-specific identifier you do not recognize (LLM model names like gpt-5.4 / claude-sonnet-4.5; host config keys like [agents] in ~/.codex/config.toml; CLI flags like --sandbox-mode; package-registry paths; service endpoints), do not flag it as fabricated based on unfamiliarity — absence of recognition is not evidence of absence. State the verification gap as a Note (cannot verify against platform docs from this reviewer's tool set — recommend external check), or skip it. You CAN flag: typos in identifiers you recognize, internal repo-relative paths that don't exist (verify with Read/Glob), and API misuses that contradict a rule you can read.
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 · 51 lines · 91 tokens per session scan A 646ee148b70e
audit-group-reviewer is an agent published in the GitHub repository CorridorTech/PoseCap (265 stars, last pushed 3d ago), licensed Apache-2.0. It adds 91 tokens to every session and 922 once invoked, about $0.0004 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 2026-08-30.
Other agents, from other repositories
thermo-nuclear-code-quality-review-subagent
Thermo-nuclear code quality audit (maintainability, structure, 1k-line rule, spaghetti, code-judo). Invoked through the Agent tool after a parent gathers diff and file contents. Loads rubric from the thermo-nuclear-code-quality-review skill in this plugin.
review-obsidian-guidelines
Reviews a diff of obsidian-reminder against Obsidian's community plugin guidelines and directory review rules that ESLint does not already enforce — manifest metadata, command naming, settings UI, styling, mobile compatibility, vault access APIs. Used by the pr-review workflow; read-only.
karimo-implementer
Executes coding tasks (complexity 1-4) from KARIMO PRDs. Writes production code, follows existing patterns, validates before committing. Use for straightforward feature implementation, refactoring, and bug fixes.
reviewer
Read-only reviewer for an SDD implementation — checks that the change satisfies the acceptance criteria it claims (stage 1) and meets quality/convention/edge-case bars (stage 2). Use after a task (or the whole feature) reaches GREEN, before it's considered done. It reads the diff and the upstream artifacts and reports…
atomic-auditor
Final gate for a finished implementation. Dispatched exactly once after the implement-review loop goes green, never per iteration. Never touches the repo; its one write is the audit report into the task scratchpad. Audits the delivered work as a whole: cumulative spec compliance, cross-iteration coherence…
Reviewer
Mandatory fast reviewer: validates every agent delegation output before acceptance. Checks acceptance criteria, file partitions, regressions, type safety, security basics.