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/vincent-k2026/codachi/claude-mdgit clone --depth 1 https://github.com/vincent-k2026/codachiWrote 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/vincent-k2026/codachi/claude-md)<a href="https://agentmods.dev/instructions/vincent-k2026/codachi/claude-md"><img src="https://agentmods.dev/badge/instructions/vincent-k2026/codachi/claude-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 | $0.01157 | $0.01157 |
| Opus 5 | $0.00579 | $0.00579 |
| Sonnet 5 | $0.00231 | $0.00231 |
| Haiku 4.5 | $0.00116 | $0.00116 |
Grade B, and why
codachi CLAUDE.md scanned grade B with 1 finding 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 5d 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.
Reads agent configuration directoriesmediumAgent snooping
.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.
node dist/index.js init # install into ~/.claude/settings.json How it starts
The opening of the file, as written. The whole thing — 62 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md
This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
Project
codachi — a tamagotchi pet that lives in the Claude Code statusline. Zero runtime dependencies, pure TypeScript compiled to ESM. Two binaries: codachi (statusline renderer, reads JSON from stdin → writes ANSI to stdout) and codachi-hook (PostToolExecution hook that records events to disk).
Commands
npm run build # tsc → dist/ + copy locales
npm run dev # tsc --watch
npm test # vitest run (full suite)
npm run test:cov # coverage
npm run bench # render p50/p95/p99 benchmark (budget 50ms)
npx vitest run src/mood.test.ts # single test file
npx vitest run -t "pattern" # single test by name
node dist/index.js demo # live terminal demo
node dist/index.js config # interactive config TUI
node dist/index.js init # install into ~/.claude/settings.json
node dist/index.js uninstall # remove from settings.json
node dist/index.js stats # local productivity dashboard
node dist/index.js plugins # list loaded plugins
Node ≥18 required. No lint step configured; tsc is the type gate.
Architecture
Data flow — two independent processes sharing state on disk:
Claude Code ──stdin JSON──▶ index.ts (renderer, p95 ~0.1ms) ──ANSI──▶ statusline
│
└──PostToolUse──▶ hook.ts ──▶ events.json ──▶ mood engine on next render
The renderer never calls APIs and never blocks on I/O longer than a git status (cached 2s). The hook's only job is to classify a tool-use event and append it; it must stay fast and cannot assume the renderer is running. Events use an optimistic-concurrency generation counter to prevent lost updates from concurrent hooks.
State lives in ~/.claude/plugins/codachi/ (session state, events, memory, error log) and ~/.config/codachi/config.json (user config). All writes are atomic (write-tmp + rename) via fs-utils.ts. Pet identity is bound to transcript_path so each Claude session gets a stable species/palette.
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.
- 5d ago First seen · 62 lines · 1,157 tokens per session scan B 0922f7907d7a
codachi CLAUDE.md is an instructions file published in the GitHub repository vincent-k2026/codachi (12 stars, last pushed 4mo ago), licensed MIT. It adds 1,157 tokens to every session, about $0.0058 per session on Opus 5. A static security scan graded it B with 1 finding (reads agent configuration directories). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
platform AGENTS.md
AGENTS.md instructions for doka-guide/platform, covering agents.md — как работать с платформой доки, 1. правила работы, сначала подумай, потом пиши, минимум кода and правки хирургические.
zapo CLAUDE.md
Claude Code instructions for vinikjkkj/zapo: The contributor handbook for this repository lives in AGENTS.md. Read that file and follow it as your primary source of project conventions, architecture rules, performance constraints, and review anti-patterns.
Silex AGENTS.md
AGENTS.md instructions for silexlabs/Silex, covering agents.md, tech stack, run it locally, building the desktop app and when writing code for silex (editing source).
SoDam-WikiMate AGENTS.md
AGENTS.md instructions for sodam-ai/SoDam-WikiMate, covering wikimate 작업 규칙 (크로스툴 — claude code · codex · gemini 공용), 문체, 자동 동작 (자연어 트리거 — 강제 호출 불필요), 접근 규칙 (자동 감지 — 특정 도구에 한정하지 않음) and obsidian 규칙.
sdk-generator AGENTS.md
AGENTS.md instructions for appwrite/sdk-generator, covering agents guide — sdk generator, critical rules for agents, rule 1: always regenerate after template edits, rule 2: new files require getfiles() registration and rule 4: copy scope = no twig processing.
hal-9000 AGENTS.md
AGENTS.md instructions for vinta/hal-9000, covering agents.md, commands, gotchas, external tool documentation and context7 library ids.