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/zhanex/legax/agents-mdgit clone --depth 1 https://github.com/zhanex/legaxWhat 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 | $0.00690 | $0.00690 |
| Opus 5 | $0.00345 | $0.00345 |
| Sonnet 5 | $0.00138 | $0.00138 |
| Haiku 4.5 | $0.00069 | $0.00069 |
Grade A, and why
legax 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 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.
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 — 50 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
English | Simplified Chinese
This is the repository instruction contract for AI coding agents working on Legax. Start here, then read Context For LLMs. Do not load every document by default.
Hard Rules
- Keep the project dependency-free: Node 18+ standard library only, no bundler, and no new npm dependency unless the design has been accepted.
- Keep configuration YAML-only. Use
scripts/lib/yaml.mjs; do not add JSON config support or a YAML dependency casually. - Keep cross-process daemon and adapter coordination in
scripts/lib/runtime-state.mjs; do not create sibling state files for cursors, selected sessions, modes, or launch requests. - Keep MCP as a capability layer. MCP tools may notify, poll, or request permission, but must not start or stop adapter processes.
- Never simulate UI clicks, auto-approve native prompts, or bypass an agent's security policy.
- Do not route remote workflow input to arbitrary shell,
eval, script, prompt, command, or free-form executable args. - Keep documentation language pairs together: English
.mdwith.zh-CN.md, andconfig.example*.yamlwith matchingconfig.example*.zh-CN.yaml. - Keep tracked documentation and commented examples UTF-8 without BOM; do not rewrite them through locale-dependent shell encodings.
- Do not commit implementation plans, drafts, scratch notes, audit notes, or agent transcripts as tracked docs. Use ignored process directories or external task systems.
- If you add a script or E2E file, append it to the explicit
package.jsoncheck:nodeortest:e2elist.
Context Loading
- Read docs/context_for_llms.md for the task routing table and canonical owners.
- Default to English docs. Load
.zh-CNpairs only for translation, Chinese documentation work, or pair consistency checks. - English docs are the canonical contract. Chinese docs are localized mirrors; if they diverge, follow English and update both.
- For non-trivial code changes, use the smallest task-specific context set from
docs/context_for_llms.md, then usedocs/CHANGE_MATRIX.mdfor focused verification. - If docs conflict with this file, stop and surface the conflict before editing.
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 · 50 lines · 690 tokens per session scan A e6b32b69c3f8
legax AGENTS.md is an instructions file published in the GitHub repository zhanex/legax (11 stars, last pushed 2mo ago), licensed MIT. It adds 690 tokens to every session, about $0.0034 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
mcp-documentation-server copilot-instructions.md
Instructions for andrea9293/mcp-documentation-server, covering mcp documentation server - ai coding agent instructions, architecture overview, core components, server entry point (src/server.ts) and embedding providers (src/embedding-provider.ts).
dotbot AGENTS.md
Instructions for andresharpe/dotbot, covering agents.md, project overview, commands, architecture and three core systems.
gemini-mcp CLAUDE.md
Claude Code instructions for RLabs-Inc/gemini-mcp, covering gemini mcp server project guide, project overview, key components, tools implemented and environment variables.
video-research-mcp CLAUDE.md
Instructions for Galbaz1/video-research-mcp, covering claude.md, memory source guard, what this is, commands and video-research-mcp (root).
video-research-mcp AGENTS.md
Instructions for Galbaz1/video-research-mcp, covering agents.md, scope, what this is, commands and code review trigger protocol.
bestprice-mcp GEMINI.md
Gemini CLI instructions for TheBestCo/bestprice-mcp: Use the BestPrice tools when a user wants to find or compare products sold in Greece or understand whether a current price is low.