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/grabrick/clave-cli/claude-mdgit clone --depth 1 https://github.com/grabrick/clave-cliWhat 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 | $0.02551 | $0.02551 |
| Opus 5 | $0.01275 | $0.01275 |
| Sonnet 5 | $0.00510 | $0.00510 |
| Haiku 4.5 | $0.00255 | $0.00255 |
Grade A, and why
clave-cli CLAUDE.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.
How it starts
The opening of the file, as written. The whole thing — 170 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md — навигатор по проекту clave
Этот файл — первое, что читает агент. Здесь собраны НЕОЧЕВИДНЫЕ тонкости и грабли, на которых легко сломать проект. Структуру кода агент увидит сам — тут то, чего в коде не видно.
0. Что такое clave
Локальный TUI-оркестратор на Rust (v0.3.0, MIT, edition 2021). Связывает CLI
claude и codex как агентов в один цикл. Зависимости минимальны: crossterm 0.29,
ratatui 0.30, serde_json, unicode-width, на unix — libc. Никаких async-рантаймов.
Режимы: чат, план (встроенный движок spec-clave вшит в бинарь), тандем
(Executor=Claude + Critic=Codex спорят до консенсуса → исполнение → ревью → правка).
1. Жёсткие правила общения и авторства
- Общение и комментарии/коммиты — на русском.
- НИКАКИХ следов ИИ. В commit-сообщениях, описаниях PR, коде и комментариях НЕ
упоминать Claude/Anthropic/ИИ, не добавлять
Co-Authored-By, не ставить футер «Generated with…». Весь код и коммиты оформляются как код владельца. - Следовать лучшим практикам, не плодить уязвимости.
2. Тулчейн (иначе cargo: command not found)
cargo/rustc лежат в ~/.cargo/bin и НЕ в PATH сессии. В начале любой команды:
export PATH="$HOME/.cargo/bin:$PATH"
3. ГЕЙТ: без него ничего не отгружается
Перед КАЖДОЙ отгрузкой прогнать ВСЁ и убедиться, что зелено. exit 0 — НЕ доказательство:
смотреть на строку test result: ok. N passed, а не на код возврата.
# PATH 1 — обычный
cargo test --bin clave
# PATH 2 — «CI без провайдеров»: пустое окружение без claude/codex в PATH.
# Ловит скрытые зависимости от установленных CLI. ОБЯЗАТЕЛЕН — на нём падало то,
# что PATH 1 пропускал.
env -i PATH="$HOME/.cargo/bin:/usr/bin:/bin:/usr/sbin:/sbin" HOME="$HOME" cargo test
cargo clippy --bin clave --all-targets # проект живёт на -D warnings, любое = красный
cargo fmt --check # `fmt --check | tail` ВРЁТ про exit — смотри вывод
4. Две ветки — САМОЕ КРИТИЧНОЕ
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 · 170 lines · 2,551 tokens per session scan A e340fc8c03c7
clave-cli CLAUDE.md is an instructions file published in the GitHub repository grabrick/clave-cli (1 stars, last pushed 1mo ago), licensed MIT. It adds 2,551 tokens to every session, about $0.0128 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 2026-08-31.
Other instructions, from other repositories
codexRS AGENTS.md
Instructions for Kiwunaka/codexRS, covering project rules, proportional engineering and universal safety.
kimi-responses-adapter AGENTS.md
AGENTS.md instructions for jianyun8023/kimi-responses-adapter, covering repository guidelines, project structure & module organization, build, test, and development commands, coding style & naming conventions and testing guidelines.
codex-session-tui AGENTS.md
Instructions for avikalpa/codex-session-tui, covering agents.md, project purpose, current scope, architecture and safety rules.
ai-history CLAUDE.md
Instructions for jiantao88/ai-history, covering claude.md, project overview, development commands, architecture and key design decisions.
cli AGENTS.md
Instructions for googleworkspace/cli, covering agents.md, project overview, build & test, changesets and architecture.
turso AGENTS.md
Instructions for tursodatabase/turso, covering turso agent guidelines, quick reference, testing, running tests and test organization.