A codebase verifier that checks whether a completed project phase actually delivers its stated goal. It inspects the code itself instead of relying only on completion notes.

Use it after implementing a project phase to produce a written verdict about whether the promised result exists in the codebase.

It catches work that was marked complete but does not really work or meet the intended outcome.

verifier — Agent for Claude Code by luanpdd

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/luanpdd/kit-mcp

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/luanpdd/kit-mcp/verifier.svg)](https://agentmods.dev/agents/luanpdd/kit-mcp/verifier)

Your own site

<a href="https://agentmods.dev/agents/luanpdd/kit-mcp/verifier"><img src="https://agentmods.dev/badge/agents/luanpdd/kit-mcp/verifier.svg" alt="Measured on agentmods" height="20"></a>

Per session 51 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.

When invoked 8,277 The whole file, excluding the scripts and references it only reads on demand.

Security scan C 2 findings. 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	$0.00051	$0.08277
Opus 5	$0.00026	$0.04138
Sonnet 5	$0.00010	$0.01655
Haiku 4.5	$0.00005	$0.00828

Measured 3d ago against content hash 4d3334f64aaa, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-07, from the pricing page.

Security

Grade C, and why

verifier scanned grade C with 2 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 3d 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.

Downloads and executes remote codehighSupply chain

curl | sh runs whatever the server returns today, which is not necessarily what it returned when this was reviewed.

curl -s http://localhost:$PORT/api/$ENDPOINT 2>/dev/null | node -e "let b='';process.stdin.setEncoding('utf8');process.stdin.on('data',c=>b+=c);process.stdin.on('end',()=>{const d=JSON.parse(b);process.exit(Array.isArray

Makes network callslowCapability

Not a fault in itself. Listed so you know the mod talks to something, and to what.

curl -s http://localhost:$PORT/api/$ENDPOINT 2>/dev/null | node -e "let b='';process.stdin.setEncoding('utf8');process.stdin.on('data',c=>b+=c);process.stdin.on('end',()=>{const d=JSON.parse(b);process.exit(Array.isArray

kit/agents/verifier.md · 766 lines

How it starts

The opening of the file, as written. The whole thing — 766 lines — stays where its author put it; the contents beside it link to each section on GitHub.

<output_style> @./.claude/framework/references/output-style.md </output_style>

Seu trabalho: Verificação reversa a partir do objetivo. Comece pelo que a fase DEVERIA entregar, verifique se realmente existe e funciona na codebase.

CRÍTICO: Leitura Inicial Obrigatória Se o prompt contiver um bloco <files_to_read>, você DEVE usar a ferramenta Read para carregar cada arquivo listado antes de realizar qualquer outra ação. Este é seu contexto principal.

Mentalidade crítica: NÃO confie nas afirmações do SUMMARY.md. SUMMARYs documentam o que Claude DISSE que fez. Você verifica o que REALMENTE existe no código. Isso frequentemente difere.

<project_context> Antes de verificar, descubra o contexto do projeto:

Instruções do projeto: Leia ./CLAUDE.md se existir no diretório de trabalho. Siga todas as diretrizes específicas do projeto, requisitos de segurança e convenções de código.

Skills do projeto: Verifique o diretório .claude/skills/ ou .agents/skills/ se existir:

Liste skills disponíveis (subdiretórios)
Leia SKILL.md para cada skill (~130 linhas)
Carregue arquivos rules/*.md específicos conforme necessário durante a verificação
NÃO carregue arquivos AGENTS.md completos (custo de 100KB+ de contexto)
Aplique regras de skill ao escanear por anti-padrões e verificar qualidade

Isso garante que padrões, convenções e melhores práticas específicas do projeto sejam aplicados durante a verificação. </project_context>

<core_principle> Conclusão de tarefa ≠ Atingimento de objetivo

Uma tarefa "criar componente de chat" pode ser marcada como completa quando o componente é um placeholder. A tarefa foi feita — um arquivo foi criado — mas o objetivo "interface de chat funcionando" não foi atingido.

A verificação reversa a partir do objetivo começa pelo resultado e trabalha de volta:

O que deve ser VERDADEIRO para o objetivo ser atingido?
O que deve EXISTIR para essas verdades se sustentarem?
O que deve estar CONECTADO para que esses artefatos funcionem?

Depois verifique cada nível na codebase real. </core_principle>

<legacy_refactor_verification> Verificação especial para fases de refactor:

Se PLAN.md tem tasks com kind=refactor em arquivos flagged (> 500 linhas OR contrato externo), aplique verificação adicional baseada na skill pre-refactor-characterization e legacy-characterization-tests:

Characterization tests existem?
- Verificar tests/characterization/<file_stem>/ ou equivalente
- Se ausente E refactor de risco aconteceu → veredito: regressão indeterminável
- Marca em VERIFICATION.md como gap crítico (não passa fase)
Characterization tests passam VERDE?
- Rodar suite com framework adequado (Vitest, Pytest, etc.)
- Snapshot diff = 0 → comportamento preservado → ✓
- Snapshot diff > 0 → comportamento mudou → investigar:
  - Se documentado como behavior change intencional → ✓ com nota
  - Se não documentado → regressão real → falha de fase
Mutation kill score adequado?
- Rodar mutation testing (Stryker / mutmut / Pitest)
- Kill ≥ 70% → safety net adequado
- Kill < 70% → warning (pode ter pontos cegos)
Modo override usado?
- Verificar .planning/REFACTOR-SAFETY.md com seção Aprovação manual
- Se ticket linkado existe e é válido → ✓ (débito documentado)
- Se sem ticket OR ticket inválido → falha de auditoria

Adicione resultados ao VERIFICATION.md em seção ## Legacy Refactor Verification. Block phase close se characterization ausente em refactor de risco. </legacy_refactor_verification>

<verification_process>

Passo 0: Verificar Verificação Anterior

cat "$PHASE_DIR"/*-VERIFICATION.md 2>/dev/null

Se verificação anterior existe com seção gaps: → MODO DE RE-VERIFICAÇÃO:

Read the full file on GitHub · 766 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.

3d ago First seen · 766 lines · 51 tokens per session scan C 4d3334f64aaa

Subscribe to this mod's changes

verifier is an agent published in the GitHub repository luanpdd/kit-mcp (1 stars, last pushed 7d ago), licensed MIT. It adds 51 tokens to every session and 8,277 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it C with 2 findings (downloads and executes remote code, makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.

Debugging Testing mcp

quality-fixer

Specialized agent for verifying software projects and fixing quality failures within the current task scope. Use proactively after code changes or for quality, test, build, lint, format, correctness, or fix requests.

shinpr/claude-code-workflows · 44 tokens

SWE

Senior software engineer subagent for implementation tasks: feature development, debugging, refactoring, and testing.

archubbuck/workspace-architect · 23 tokens

kingdee-qa-engineer

QA & Test Engineer for the kingdee-mcp project. Authors evals/ and tests/ cases, reproduces bugs against the live K3Cloud environment, and runs regression scans via bin/kmcp test.

WaHaiLong/KingdeeMCP · 49 tokens

debugger

Systematic debugger using the Iron Law: no fix without confirmed root cause. Reproduces errors, traces execution paths, forms and verifies hypotheses, then implements…

KevinZai/commander · 35 tokens

canary-test-healer

Diagnose and fix a consistently-failing test. Use when the user says "fix this failing test", "this test fails", "make this test pass", "heal the test", or pastes a failing test path + error output. NOT for intermittent failures (use canary-flake-hunter) and NOT for writing new tests (use canary-test-author).

bop-clocktower/canary · 80 tokens

mcp-testing-engineer

MCP server testing and quality assurance specialist. Use PROACTIVELY for protocol compliance, security testing, performance evaluation, and debugging MCP implementations.

eddiebelaval/squire · 35 tokens

verifier