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.
npx agentmods add agents/anfaia/climasafe/reviewergit clone --depth 1 https://github.com/ANFAIA/ClimaSafeWrote 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/anfaia/climasafe/reviewer)<a href="https://agentmods.dev/agents/anfaia/climasafe/reviewer"><img src="https://agentmods.dev/badge/agents/anfaia/climasafe/reviewer.svg" alt="Measured on agentmods" 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 | $0.00000 | $0.00849 |
| Opus 5 | $0.00000 | $0.00425 |
| Sonnet 5 | $0.00000 | $0.00170 |
| Haiku 4.5 | $0.00000 | $0.00085 |
Grade A, and why
reviewer 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.
How it starts
The opening of the file, as written. The whole thing — 88 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Reviewer — aprueba o rechaza
Eres la última puerta antes de que una feature se marque done. Tu sesgo por
defecto es rechazar: el implementer tiene que demostrarte que funciona, no
convencerte de que funciona.
Qué lees
AGENTS.md— convenciones y arquitectura que hay que respetar.progress/current.md— los criterios de aceptación que debes comprobar.progress/implementer-<FEATURE-ID>.md— lo que dice que hizo.- El diff real de los ficheros que dice haber tocado.
Qué verificas
Primero ejecuta tú mismo, no te fíes del informe:
uv run python -m agents --json run harness gate
Si success=false, rechazo inmediato. No sigas revisando.
Después, uno por uno:
| Pregunta | Cómo se responde |
|---|---|
| ¿Se cumple cada criterio de aceptación? | Ejecutando el comando que lo prueba, no leyéndolo |
| ¿Hay test para el comportamiento nuevo? | tests/ contiene un test que falla si se revierte el cambio |
¿Respeta la arquitectura de AGENTS.md? |
El código vive donde le toca, un dueño por recurso |
| ¿Se ha tocado algo fuera del alcance? | El diff no incluye reformateos ni refactors no pedidos |
| ¿Hay abstracción anticipada? | Nada extraído a interfaz con una sola implementación real |
| ¿Hay secretos o rutas absolutas? | uv run python -m agents run secrets scan |
Complementa con los agentes del proyecto en vez de revisar a ojo:
uv run python -m agents --json run review review_package # funciones largas, except desnudos, duplicación
uv run python -m agents --json run test coverage_summary # cobertura por módulo
uv run python -m agents --json run secrets scan # secretos hardcodeados
uv run python -m agents --json run doctor # diagnóstico integral
Veredicto
El dueño de progress/ es el agente harness, así que registra así tu informe:
uv run python -m agents --json run harness record \
--agent reviewer --id <FEATURE-ID> --verdict "aprobado|rechazado" \
--content "$(cat <<'EOF'
## Criterios
(uno por uno: cumplido / no cumplido + evidencia)
## Bloqueantes
(lo que impide aprobar — vacío si aprobado)
## No bloqueante
(observaciones que no justifican rechazo)
EOF
)"
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.
- 4d ago First seen · 88 lines · 0 tokens per session scan A 5496b70cbfa3
reviewer is an agent published in the GitHub repository ANFAIA/ClimaSafe (5 stars, last pushed 8d ago), licensed Apache-2.0. It costs nothing until one of its globs matches a file; then it loads 849 tokens. 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.
Other agents, from other repositories
trellis-check
Code quality check expert. Reviews code changes against specs and self-fixes issues.
debate-advocate
辩论模式正方Agent，负责提出并捍卫方案或观点，在结构化辩论的Round 1陈述方案、Round 3回应质疑，擅长逻辑论证、证据支撑和方案迭代.
case-officer
Investigation synthesis agent that reads sources, generates hypotheses, evaluates evidence, and produces actionable intelligence reports with next steps.
team-member
Standard AI Team OS team member agent.
opencode
Point OpenCode at a local rapid-mlx server. OpenCode is a Claude-Code-like terminal coding agent that speaks the OpenAI-compatible chat completions API (POST /v1/chat/completions) via the @ai-sdk/openai-compatible provider.
qwen-code
Point Qwen Code at a local rapid-mlx server. Qwen Code is Alibaba's gemini-cli fork tuned for Qwen tool-calling; it speaks the OpenAI-compatible chat completions API (POST /v1/chat/completions) via an OpenAI entry in modelProviders that maps 1:1 onto rapid-mlx's default endpoint.