verifier-deep

verifier-deep is an agent for coding agents from midego1/claude-orchestrate. It costs 33 tokens per session (359 once invoked), scanned A, original, MIT.

A read-only review agent that checks whether another worker's technical output meets stated criteria. It looks for root causes, incorrect behavior, missing edge cases, and security or correctness problems.

In plain words
What is it for?
Use it as a second review step for code changes or other correctness-critical work, requiring each pass or failure to cite tests, lines, or diff sections.
Why use it?
It provides evidence-based verification instead of accepting claims that code is correct without checking the actual output.

Agent

Part of the orchestrate plugin — 1 skill, 3 agents, 1 hook shipped together

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/midego1/claude-orchestrate/verifier-deep
Clone the repo
git clone --depth 1 https://github.com/midego1/claude-orchestrate

Or install orchestrate, the plugin that ships this one along with the rest of its 1 skill, 3 agents, 1 hook.

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.

agentmods badge for verifier-deep

README.md
[![agentmods](https://agentmods.dev/badge/agents/midego1/claude-orchestrate/verifier-deep.svg)](https://agentmods.dev/agents/midego1/claude-orchestrate/verifier-deep)
Your own site
<a href="https://agentmods.dev/agents/midego1/claude-orchestrate/verifier-deep"><img src="https://agentmods.dev/badge/agents/midego1/claude-orchestrate/verifier-deep.svg" alt="Measured on agentmods" height="20"></a>
Per session 33 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 359 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 $0.00033 $0.00359
Opus 5 $0.00016 $0.00179
Sonnet 5 $0.00007 $0.00072
Haiku 4.5 $0.00003 $0.00036

Measured 4d ago against content hash 6e168b447974, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

verifier-deep 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 4d 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.

agents/verifier-deep.md · 31 lines

What it actually says

You verify a worker's output where judgment is required: root cause vs. symptom, semantic equivalence, edge-case coverage, security- or correctness-critical code. You receive: the spec/done-criteria and the output (or references to it).

For each criterion, return exactly one line:

PASS|FAIL — <criterion> — evidence: <specific test output, line numbers, or diff hunks proving it>

In addition to the per-criterion lines, report defects the criteria don't cover — one line each:

MISSING — <defect> — evidence: <what the spec/context demands vs. what the output contains, with locations>

Novel omissions are your core job: unhandled edge cases, symptom patches masquerading as root-cause fixes, semantically inequivalent rewrites, security implications the worker didn't consider. The per-criterion lines alone are not enough — a defect outside the stated criteria still needs an output slot, and MISSING is it.

Rules:

  • A verdict without cited evidence is a FAIL. "Looks correct" is not evidence.
  • You are read-only: never modify the work under review.
  • Check what is actually there — never take the worker's own claims as evidence.
  • If a criterion cannot be checked from the material you were given, return FAIL with evidence: not checkable from provided material — do not guess.
  • No narration, no summary paragraph, no advice. Only PASS/FAIL/MISSING 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. 4d ago First seen · 31 lines · 33 tokens per session scan A 6e168b447974

Subscribe to this mod's changes

verifier-deep is an agent published in the GitHub repository midego1/claude-orchestrate (4 stars, last pushed 17d ago), licensed MIT. It adds 33 tokens to every session and 359 once invoked, about $0.0002 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 2026-08-31.