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 instructions/anfaia/climasafe/agents-mdgit 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/instructions/anfaia/climasafe/agents-md)<a href="https://agentmods.dev/instructions/anfaia/climasafe/agents-md"><img src="https://agentmods.dev/badge/instructions/anfaia/climasafe/agents-md.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 | $0.02661 | $0.02661 |
| Opus 5 | $0.01331 | $0.01331 |
| Sonnet 5 | $0.00532 | $0.00532 |
| Haiku 4.5 | $0.00266 | $0.00266 |
Grade A, and why
ClimaSafe AGENTS.md 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 — 249 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Sistema de Agentes — ClimaSafeAI
Este proyecto incluye un sistema de 26 agentes que automatizan todo el ciclo de desarrollo: desde análisis de datos hasta release, pasando por revisión de código, tests, dependencias y despliegue.
Git en este repo: los commits los decide el humano y los hace el agente
git— Conventional Commits, sin línea de co-autoría ni pie de ninguna herramienta de IA:uv run python -m agents --json run git suggest_commit_message # propone el mensaje uv run python -m agents --json run git commit_with_changelog --message "<msg>"Con una excepción automática: al cerrar una feature, el líder pasa
--commitaharness finishy entonces sí commitea por sí solo solo las rutas de ese ticket (las de--changesmás los ficheros del propio cierre: featureslist.json, progress/, pyproject.toml, README.md). Sin ese flag,finishpropone el mensaje y NO commitea. El resto del árbol queda sin commitear y lo decide el humano. Ningún otro asistente ejecutagitnighpor su cuenta ni por iniciativa propia — ni siquiera para aplicar o deshacer un cambio.
Protocolo del arnés — LEE ESTO PRIMERO
Este fichero es el punto de entrada. Todo agente que trabaje en este proyecto sigue estos pasos en orden, antes de escribir una sola línea de código.
1. ./init.sh ¿el entorno está sano? si no → PARA
2. progress/current.md ¿hay trabajo a medias? si sí → retómalo
3. featureslist.json primera feature pendiente con deps en done
4. marcar in_progress en featureslist.json + rellenar current.md
5. delegar explorer → implementer → reviewer
6. verificar ./init.sh en verde + criterios uno a uno
7. done featureslist.json + resumen en history.md
La regla que no se salta
Ninguna feature se marca done sin que ./init.sh pase en verde.
Y no es una instrucción: es código. harness finish ejecuta la puerta antes de
tocar el backlog y devuelve success=false si está en rojo o si no le pasas
evidencia. No hay forma de rodearlo pidiéndoselo amablemente al modelo — la
única sería editar el JSON a mano, y eso está prohibido explícitamente.
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 · 249 lines · 2,661 tokens per session scan A 39f1aa1f6f5d
ClimaSafe AGENTS.md is an instructions file published in the GitHub repository ANFAIA/ClimaSafe (5 stars, last pushed 7d ago), licensed Apache-2.0. It adds 2,661 tokens to every session, about $0.0133 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.
Other instructions, from other repositories
Kokoro-FastAPI AGENTS.md
AGENTS.md instructions for remsky/Kokoro-FastAPI, covering agents.md, layout, commands, conventions and gotchas.
geolens CLAUDE.md
Claude Code instructions for geolens-io/geolens: This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
substack-gateway-oss CLAUDE.md
Claude Code instructions for jakub-k-slys/substack-gateway-oss, covering claude.md, commands, install dependencies (dev included, all workspace members), run the server (dev mode with reload) and lint.
Huntable-CTI-Studio AGENTS.md
AGENTS.md instructions for dfirtnt/Huntable-CTI-Studio, covering agents.md, orientation, local context, prompt-injection alerting and change-type quick reference.
deckforge CLAUDE.md
Instructions for Whatsonyourmind/deckforge, covering deckforge, tech stack, project structure, key commands and local development.
dev-challenge CLAUDE.md
Claude Code instructions for micheltlutz/dev-challenge, covering claude.md, the one rule to internalise, slash commands, subagents and skills.