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/gsa-tts/agentic-coding-playbook/agents-mdgit clone --depth 1 https://github.com/GSA-TTS/agentic-coding-playbookWrote 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/gsa-tts/agentic-coding-playbook/agents-md)<a href="https://agentmods.dev/instructions/gsa-tts/agentic-coding-playbook/agents-md"><img src="https://agentmods.dev/badge/instructions/gsa-tts/agentic-coding-playbook/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 | $0.10717 | $0.10717 |
| Opus 5 | $0.05359 | $0.05359 |
| Sonnet 5 | $0.02143 | $0.02143 |
| Haiku 4.5 | $0.01072 | $0.01072 |
Grade F, and why
agentic-coding-playbook AGENTS.md scanned grade F with 3 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.
Hidden instructionshighPrompt injection
Directives inside HTML comments, invisible characters or bidirectional overrides are read by the model and not by the person reviewing the file.
<!-- NIST SP 800-53: SA-5 (System Documentation), SA-8 (Security Engineering Principles), SA-15 (Development Process), SA-17 (Developer Security Architecture) --> Cloud metadata endpointhighServer-side request forgery
One request to 169.254.169.254 can return temporary IAM credentials.
- Access cloud metadata endpoints (169.254.169.254, etc.) unless specifically authorized Unrestricted tool accessmediumExcessive agency
A wildcard tool grant or "run any command" leaves no least-privilege boundary at all.
| Execute arbitrary code from external sources | Prevents remote code execution attacks | SI-3, CM-7 | How it starts
The opening of the file, as written. The whole thing — 863 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md — Federal AI Agent Behavioral Best Practices
Version: 1.0.0 | Impact Level: FIPS Moderate | Scope: Single-agent, internal enterprise
Quick Reference
| Rule | Requirement |
|---|---|
| Priority | safety > correctness > compliance > simplicity > performance |
| Identity | Document AI usage (PR-level recommended, commit-level optional), log all actions, identify as AI when asked |
| Permissions | Explicit allowlist — only permitted actions without approval |
| Prohibited | No secrets in code, no eval/exec with external data, no production DB access |
| Data | Field-level encryption for PII, mask in logs, secrets from approved KMS only |
| Network | TLS 1.2+, explicit allowlist, no unapproved outbound connections |
| Dependencies | Pin exact versions, verify names (typosquatting), check CVEs, check licenses |
| Testing | Unit + integration tests required, all must pass before declaring done |
| Changes | Plan before execute, PR with verification transcript, no silent failures |
| Engineering | ADR for architecture changes, flag size/complexity violations, docs-as-code |
Full details in sections below. See
CONTEXT-GUIDE.mdfor loading instructions.
Disclaimer: This playbook provides informational best practices — it is not official GSA policy or authoritative federal guidance. Each agency must tailor these recommendations to their specific ATO requirements, organizational policies, and risk tolerance. This content supports the Agentic Coding Capability Assessment being conducted at GSA-TTS.
This document defines the behavioral best practices that AI coding agents SHOULD follow when assisting federal employees with software development. Place this file (or a customized copy from templates/AGENTS.md.template) in the root of your repository.
Key words: "MUST", "MUST NOT", "SHOULD", "SHOULD NOT", and "MAY" are used per RFC 2119.
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 · 863 lines · 10,717 tokens per session scan F 578d2f942979
agentic-coding-playbook AGENTS.md is an instructions file published in the GitHub repository GSA-TTS/agentic-coding-playbook (21 stars, last pushed 7d ago), licensed CC0-1.0. It adds 10,717 tokens to every session, about $0.0536 per session on Opus 5. A static security scan graded it F with 3 findings (hidden instructions, cloud metadata endpoint, unrestricted tool access). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
aguara CLAUDE.md
Claude Code instructions for garagon/aguara, covering claude.md, project status, build & test commands, single package test and single test function.
everything-ai-coding AGENTS.md
Instructions for zgsm-ai/everything-ai-coding, covering agents.md, project context, do not commit openspec/, test files and commit conventions.
aigis CLAUDE.md
Instructions for killertcell428/aigis, covering aigis — project guidelines, zenn記事管理（zenn cli + github連携）, ディレクトリ構成, 記事の作成 and プレビュー.
zod-skill AGENTS.md
Instructions for anivar/zod-skill, covering zod v4 — complete guide, abstract, 1. parsing & type safety, rule: use safeparse() for user input and rule: use parseasync for async refinements.
claudini CLAUDE.md
Instructions for romovpa/claudini, covering claudini — developer guide, package manager, running benchmarks, run a preset (all methods, all samples) and override method, sample, seed, budget from cli.
tooltrust-scanner CLAUDE.md
Instructions for AgentSafe-AI/tooltrust-scanner: Use the /browse skill from gstack for all web browsing. Never use mcpclaude-in-chrome tools.