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/agent-sh/agnix/agents-mdgit clone --depth 1 https://github.com/agent-sh/agnixWhat 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 | $0.02708 | $0.02708 |
| Opus 5 | $0.01354 | $0.01354 |
| Sonnet 5 | $0.00542 | $0.00542 |
| Haiku 4.5 | $0.00271 | $0.00271 |
Grade A, and why
agnix 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 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.
Copies of this mod
1 near-identical copy found in the catalogue:
- agnix CLAUDE.md — 100% identical, 0 lines differ
How it starts
The opening of the file, as written. The whole thing — 225 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Project Memory: agnix
Linter for agent configurations. Validates Skills, Hooks, MCP, Memory, Plugins.
Repository: https://github.com/agent-sh/agnix
Project Instruction Files
CLAUDE.mdis the project memory entrypoint for Claude Code.AGENTS.mdis a byte-for-byte copy ofCLAUDE.mdfor tools that readAGENTS.md(Codex CLI, OpenCode, Cursor, Cline, Copilot).- Keep them identical (tests enforce this).
Critical Rules
- Rust workspace - agnix-rules (data), agnix-core (lib), agnix-cli/agnix-lsp/agnix-mcp (binaries), agnix-wasm (WASM bindings)
- rules.json is source of truth -
knowledge-base/rules.jsonis the machine-readable source of truth. When adding a new rule, add it to BOTHrules.jsonANDVALIDATION-RULES.md. CI parity tests enforce this. - Plain text output - No emojis, no ASCII art
- Certainty filtering - HIGH (>95%), MEDIUM (75-95%), LOW (<75%)
- Release binaries - Compile with LTO, strip symbols
- Track work in GitHub issues - All tasks tracked there
- Task is not done until tests added - Every feature/fix must have quality tests
- Documentation - Keep long-form docs in
README.md,SPEC.md, andknowledge-base/(especiallyknowledge-base/VALIDATION-RULES.md). KeepCLAUDE.md/AGENTS.mdfor agent instructions only. - Always follow the skill/command flow as instructed - No deviations
- No unnecessary files - Don't create summary files, plan files, or temp docs unless specifically required
- Never merge without waiting for the
revuto-reviewcheck to end successfully - It might take time, but this is the major quality gate and most thorough review. - You MUST follow the flow phases one by one - If they state to use subagents, tools, or any specific method, you must follow it exactly as described.
- You MUST address all comments and reviews - If reviewers leave comments, even minor ones, and even if not a requested change, you must address them all before merging. If you disagree, respond in the review comments. Minor comments must still be addressed.
- Use single dash for em-dashes - In prose, use
-(single dash with spaces), never--(double dash). This does not apply to CLI flags like--helpor--fix.
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.
- 2d ago First seen · 225 lines · 2,708 tokens per session scan A 86169395717e
agnix AGENTS.md is an instructions file published in the GitHub repository agent-sh/agnix (400 stars, last pushed 3d ago), licensed Apache-2.0. It adds 2,708 tokens to every session, about $0.0135 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 2026-08-30.
Other instructions, from other repositories
seek AGENTS.md
Instructions for whyiyhw/seek, covering seek — project guide for ai assistants, pitfall recording (load-bearing — read this), architecture (non-negotiable), tool usage workflow (load-bearing) and tool descriptions: the highest-leverage behavioural lever.
building-a-coding-agent-from-scratch-course AGENTS.md
Instructions for decodingai-magazine/building-a-coding-agent-from-scratch-course, covering decode, project structure, tech stack, docs & external services and running commands.
cyrus copilot-instructions.md
Copilot instructions for cyrusagents/cyrus: Note, there is a need to maintain the use of '--print' when running the claude exec commands because that is what makes it non-interactive.
framework AGENTS.md
Instructions for ai-driven-dev/framework, covering agents.md, behavior, communication, action and memory management.
spellbook AGENTS.md
Instructions for majiayu000/spellbook, covering spellbook agent contract, routing, scope rules, threads long-run guardrails and validation.
skills CLAUDE.md
Instructions for capawesome-team/skills, covering claude.md, project overview, repository structure, skill.md format and writing agent-optimized skills.