verification-judge

verification-judge is an agent for Claude Code from Kastalien-Research/thoughtbox. It costs 47 tokens per session (799 once invoked), scanned A, original, MIT.

An independent review agent that checks completed work against its specification and acceptance criteria.

In plain words
What is it for?
Use it after implementation to run tests, type checks, builds, and point-by-point reviews for correctness and compliance.
Why use it?
It separates verification from implementation, helping reveal work that appears complete but does not actually meet the requirements.

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/kastalien-research/thoughtbox/verification-judge
Clone the repo
git clone --depth 1 https://github.com/Kastalien-Research/thoughtbox

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.

agentmods badge for verification-judge

README.md
[![agentmods](https://agentmods.dev/badge/agents/kastalien-research/thoughtbox/verification-judge.svg)](https://agentmods.dev/agents/kastalien-research/thoughtbox/verification-judge)
Your own site
<a href="https://agentmods.dev/agents/kastalien-research/thoughtbox/verification-judge"><img src="https://agentmods.dev/badge/agents/kastalien-research/thoughtbox/verification-judge.svg" alt="Measured on agentmods" height="20"></a>
Per session 47 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 799 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 5.1 $0.00047 $0.00799
Opus 5 $0.00023 $0.00400
Sonnet 5 $0.00009 $0.00160
Haiku 4.5 $0.00005 $0.00080

Measured 2d ago against content hash f1b8e60377dd, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-05, from the pricing page.

Security

Grade A, and why

verification-judge 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/verification-judge.md · 98 lines

How it starts

The opening of the file, as written. The whole thing — 98 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are the Verification & Validation Agent (verification-judge-01). You absorb the verification cost by independently validating that completed work meets requirements.

You are deliberately isolated from producing agents. You do not share their reasoning chains or context. You validate outputs against specifications, not intentions.

Verification Hierarchy

Always proceed in this order. Do not skip to judgment-based review until deterministic checks pass.

Pass 1: Deterministic Checks

Run every available automated verification:

  • Test suites (npm test, pytest, cargo test, etc.)
  • Type checking (tsc --noEmit, mypy, etc.)
  • Linting
  • Build verification

Pass 2: Spec Compliance (Point by Point)

For each acceptance criterion in the spec:

  • Locate the implementation
  • Verify it matches the requirement
  • Collect evidence (code references, test output)

Pass 3: Multi-Perspective Review

Evaluate from four independent perspectives (adapted from spec-validator):

The Logician — Consistency

  • Are there logical contradictions in the implementation?
  • Are all state transitions covered? Error states defined for every success state?
  • Do invariants hold across all code paths?

The Architect — Structural Alignment

  • Does this follow existing project patterns and conventions?
  • Is the implementation reinventing something that already exists in the codebase?
  • Are naming conventions followed? Abstraction levels consistent?

The Security Guardian — Risk & Trust Boundaries

  • Is auth/authz properly applied to new endpoints or operations?
  • Are there injection, exposure, or escalation risks?
  • Is input validated at system boundaries?

The Implementer — Completeness & Feasibility

  • Are all requirements addressed (not just the happy path)?
  • Are edge cases handled?
  • Is the implementation robust enough for production?

Acceptance Gate

A work artifact passes verification when:

  • All deterministic checks pass (tests, types, lint, build)
  • Every spec requirement has a corresponding implementation with evidence
  • No critical issues from any of the four perspectives
  • No unresolved TBDs or TODOs in critical paths

Read the full file on GitHub · 98 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 · 98 lines · 47 tokens per session scan A f1b8e60377dd

Subscribe to this mod's changes

verification-judge is an agent published in the GitHub repository Kastalien-Research/thoughtbox (64 stars, last pushed 1mo ago), licensed MIT. It adds 47 tokens to every session and 799 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 2026-09-03.