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/sergeiwallace/ai-cli-utils/agents-mdgit clone --depth 1 https://github.com/sergeiwallace/ai-cli-utilsWhat 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 | $0.00507 | $0.00507 |
| Opus 5 | $0.00253 | $0.00253 |
| Sonnet 5 | $0.00101 | $0.00101 |
| Haiku 4.5 | $0.00051 | $0.00051 |
Grade A, and why
ai-cli-utils 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.
How it starts
The opening of the file, as written. The whole thing — 29 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Lean configuration
Public Open-Source Package Standards
This is a public open-source package. All code, docs, comments, tests, and commit messages must be written for a general audience.
- No proprietary names — never reference ai-core, aido, or private platform/tool names in code, docs, comments, or tests.
- No personal identifiers — no personal names (first or last), usernames, private server IPs/hostnames, or account-specific paths. Use generic placeholders:
user,myproject,example.com,192.0.2.x. - No private project names or prefixes — do not hardcode personal-workflow project names or session prefixes in source, docs, or tests. Use only generic names such as
myproject,myapp,session-1, andtest-session. - Generic examples throughout — all session names, project names, and configuration values in tests/docs must be obvious placeholders, not possible real personal-workflow artifacts.
- OS portability — account for Windows, macOS, and Linux differences. Do not assume macOS-only locations or tools (for example
~/Library/,pbcopy, oropen). Usesys.platform,pathlib, andosabstractions. Flag unavoidable platform-specific code with a comment. - Commit messages are public — the same rules apply to git commit messages.
- If an existing violation is found while working, flag it immediately rather than letting it accumulate.
CLI Conventions
- All options must have both short and long forms — for example
-f/--force,-d/--dry-run. Do not add long-only flags, except hidden internalSUPPRESSflags passed machine-to-machine. - When adding a new CLI option, add both forms simultaneously. Do not ship long-only flags.
Test Requirements
Python
- Test names:
test_{given}_{when}_{then} - Use pytest fixtures for shared setup.
- A reviewer role audits test quality on every review; see
.claude/agents/reviewer.md. - Run:
pytest # pragma: no coveris a hard human gate — never add it autonomously. If a line cannot be covered, document the specific line, why it cannot be mocked, and options in the UAT report. Wait for explicit user approval before adding a pragma.
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 · 29 lines · 507 tokens per session scan A c1b46be3f26e
ai-cli-utils AGENTS.md is an instructions file published in the GitHub repository sergeiwallace/ai-cli-utils (2 stars, last pushed 2d ago), licensed MIT. It adds 507 tokens to every session, about $0.0025 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
browser-use AGENTS.md
AGENTS.md instructions for browser-use/browser-use, covering agents.md version 2, development rules, quickstart, 1. installing browser-use and on windows use .venv\scripts\activate.
browser-use CLAUDE.md
Claude Code instructions for browser-use/browser-use, covering claude.md, high-level architecture, core components, event-driven browser management and cdp integration.
agentic-context-engine AGENTS.md
Instructions for kayba-ai/agentic-context-engine, covering agents.md, repository guidelines, pipeline-first development (mandatory), core code protection and documentation maintenance.
claude-code-usage-bar AGENTS.md
Instructions for leeguooooo/claude-code-usage-bar, covering repository guidelines, project structure & module organization, build, test, and development commands, coding style & naming conventions and testing guidelines.
JIRA-Assistant-Skills CLAUDE.md
Instructions for grandcamel/JIRA-Assistant-Skills, covering jira assistant skills, project overview, architecture, directory structure and shared library pattern.
claude-tui CLAUDE.md
Instructions for slima4/claude-tui, covering claude.md, project overview, top-level scripts, tools and claude-code-statusline.