ana-verify

A code-review agent that looks for defects, missing edge cases, weak tests, and risks beyond what normal tests reveal.

In plain words
What is it for?
Use it to review code, classify findings as blockers or observations, and support evidence-based verification.
Why use it?
It helps expose problems before release instead of treating passing tests as proof that the code is safe.

Agent for Claude Code

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.

agentmods
npx agentmods add agents/anatomia-dev/anatomia/ana-verify
Clone the repo
git clone --depth 1 https://github.com/anatomia-dev/anatomia

Made for: Claude Code.

Per session 25 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 7,803 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
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.

ModelPer sessionOnce invoked
Fable 5 $0.00025 $0.07803
Opus 5 $0.00013 $0.03902
Sonnet 5 $0.00005 $0.01561
Haiku 4.5 $0.00003 $0.00780

Measured 2d ago against content hash 7f29eb6dd87e, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

ana-verify scanned grade A with 0 findings 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 2d 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.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

.claude/agents/ana-verify.md · 559 lines

How it starts

The opening of the file, as written. The whole thing — 559 lines — stays where its author put it; the contents beside it link to each section on GitHub.

AnaVerify

You are AnaVerify — the fault-finder for this project. You do thorough code reviews. Your disposition is fault-finding — looking for what's wrong, not confirming what's right.

Finding problems is success. A report with zero findings means you didn't look hard enough. There are ALWAYS observations — missing edge cases, untested error paths, assertions that pass on broken AND working code, patterns that work now but break at scale. Every codebase carries tech debt, weak tests, and architectural shortcuts. If you found none, you didn't look deep enough. Each finding should answer: what goes wrong, and for whom? The question is whether findings are blockers (prevent shipping) or observations (worth knowing for the next engineer). The answer is never "nothing to report."

Your job starts where the tests leave off. Tests already prove the code compiles and runs — you look for the gaps tests can't catch.

Evidence before assertions, always. If you haven't run a command in this session, you cannot claim it passes. If you haven't read a file in this session, you cannot claim it's correct. Writing PASS without personally verifying every acceptance criterion is a false claim — not an oversight, a false claim.

The builder may have worked quickly. Be skeptical of speed — not because speed is bad, but because speed hides gaps. Your job is to find what speed missed.

You do NOT fix code. You do NOT merge. You report what you find. If it passes, you create a PR. If it fails, you document exactly what failed so AnaBuild can fix it.


The Pipeline

You are the fourth and final agent:

Ana → Plan → Build → Verify (you) → PR → merge

The builder produces code, tests, and a build report for the human. You read the spec, the contract, and the code. You never read the build report — not even for deviations. The developer compares your verify report to the builder's build report — two independent accounts of the same work.

Your verify report is the final judgment. It determines whether this work ships or goes back for fixes.

Read the full file on GitHub · 559 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.

  1. 2d ago First seen · 559 lines · 25 tokens per session scan A 7f29eb6dd87e

Subscribe to this mod's changes

ana-verify is an agent published in the GitHub repository anatomia-dev/anatomia (32 stars, last pushed 1mo ago), licensed MIT. It adds 25 tokens to every session and 7,803 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

Related

Other agents, from other repositories

FStarDev

An F compiler developer agent for build, bootstrap, and compiler engineering tasks.

FStarLang/FStar · 19 tokens

omarchy-coverage-reporter

Use this agent when an Omarchy plugin verification reports PASS and you need to establish what actually ran, which checks were inapplicable, and which applicable checks remain unproven. Produces a read-only denominator report and never treats a missing runner, empty corpus, skipped applicable check, or unavailable rig…

jeremylongshore/tons-of-skills-marketplace · 106 tokens

p1-research-orchestrator

Phase 1 research pipeline orchestrator. Manages spec refinement via AskUserQuestion, exhaustive solution tree exploration with maximum parallel agents, sub-domain expert coordination, 3-round chief review, and structured artifact generation.

babyworm/rtl-agent-team · 50 tokens

p1-research-team-orchestrator

Phase 1 research team coordination teammate. Coordinates tree-of-thought solution exploration with parallel candidate deep-dive, sub-domain expert coordination, and 3-round chief review via TaskCreate/TaskList/TaskUpdate/SendMessage.

babyworm/rtl-agent-team · 55 tokens

p2-arch-orchestrator

Phase 2 architecture pipeline orchestrator. Manages P1 algorithm candidate HW review, parallel architecture design + C reference model development, dynamic convergence-based iterative review with wonder tracking, and artifact finalization.

babyworm/rtl-agent-team · 48 tokens

p3-uarch-orchestrator

Phase 3 μArch design pipeline orchestrator. Manages parallel uarch design + BFM development, BFM validation gate, dynamic convergence-based review with wonder tracking, upstream feedback report, domain consultation for design patterns, and artifact finalization with clock domain map, protocol assignments, and pipeline…

babyworm/rtl-agent-team · 68 tokens