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/mackding/codexclaw/agents-mdgit clone --depth 1 https://github.com/MackDing/CodexClawWhat 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 | $0.00649 | $0.00649 |
| Opus 5 | $0.00324 | $0.00324 |
| Sonnet 5 | $0.00130 | $0.00130 |
| Haiku 4.5 | $0.00065 | $0.00065 |
Grade A, and why
CodexClaw 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 yesterday.
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 — 59 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Repository Guidelines
Repo Structure
src/index.ts: application entrypoint and subsystem wiring.src/bot/: Telegram handlers, formatting, command parsing, i18n, middleware.src/orchestrator/: routing, MCP client, skill registry, GitHub/MCP skills.src/runner/: Codex SDK/CLI runner management, PTY fallback handling, and restricted shell execution.src/cron/: scheduled proactive jobs.tests/: Node built-in test suite, one*.test.jsfile per module area.
Start And Dev Commands
npm install: install dependencies.npm run start: start the Telegram bot with the current.env.npm run dev: run the bot in watch mode for local development.- Use
CODEX_BACKEND=sdkfor the preferred SDK-backed runner; keepCODEX_BACKEND=clionly when you specifically need the legacy PTY/CLI path.
Test Commands
npm test: run the full unit test suite with the Node test runner plustsx.npm run check: run the repository typecheck gate.npm run typecheck: runtsc --noEmitdirectly.npm run healthcheck: run the local runtime health check.npm run healthcheck:live: run the live Codex + Telegram health probe when local credentials are configured.
Lint And Format
npm run lint: run ESLint over source, tests, scripts, and local JS/CJS config files.npm run lint:fix: apply safe ESLint fixes.npm run format: run Prettier across the repository.npm run format:check: verify formatting without writing changes.- Do not submit formatting-only churn or rewrap unrelated files.
Files And Paths You Must Not Change
- Do not edit
.git/ornode_modules/. - Do not commit or rewrite
.env, secrets, Telegram tokens, or local session artifacts. - Do not manually edit
.codex-telegram-claws-state.json; it is runtime state. - Avoid changing files outside this repository root, even when
/repoor shell features reference other workspaces.
Contribution Rules
- Use ES Modules and keep new files under the existing feature-oriented layout.
- Keep user-facing bot copy in English by default. Localized strings must go through
src/bot/i18n.js. - Prefer focused changes. Do not mix feature work with unrelated refactors.
- Add or update tests for behavior changes in
tests/.
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.
- yesterday First seen · 59 lines · 649 tokens per session scan A c3ff6bdbc834
CodexClaw AGENTS.md is an instructions file published in the GitHub repository MackDing/CodexClaw (33 stars, last pushed 26d ago), licensed MIT. It adds 649 tokens to every session, about $0.0032 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
ai-code-interface.el AGENTS.md
Instructions for tninja/ai-code-interface.el, covering agents.md, common development commands, byte compilation, running tests and run a specific test file.
ai-code-interface.el copilot-instructions.md
Instructions for tninja/ai-code-interface.el, covering github copilot instructions for ai-code-interface.el, project overview, language and technology, code style and conventions and file naming.
bestie AGENTS.md
Instructions for sirquy/bestie, covering ai coding agent instructions, project state, current scope, architecture rules and safety and privacy.
codex-usage-widget AGENTS.md
Instructions for ognjeeen/codex-usage-widget, covering agents.md, project, scope and decision discipline, authorization and repository safety and architecture boundaries.
codex-token-meter AGENTS.md
Instructions for prefect12/codex-token-meter, covering ai token meter development guide, project shape, build and verification, git workflow and ai token meter test loop.
codex-loop-orchestra AGENTS.md
Instructions for LEO001020/codex-loop-orchestra, covering loop discipline, 1. decomposition discipline (single-pass plan-and-solve), 3. return convention (all subagents, mandatory), 4. recoverable compression directive and 5. kernel trigger rules (ipybox, when enabled).