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/vibgrate/cli/agents-mdgit clone --depth 1 https://github.com/vibgrate/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 | $0.00953 | $0.00953 |
| Opus 5 | $0.00477 | $0.00477 |
| Sonnet 5 | $0.00191 | $0.00191 |
| Haiku 4.5 | $0.00095 | $0.00095 |
Grade A, and why
cli 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 — 88 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
Guidance for AI coding agents working in this repository (the public, Apache-2.0 Vibgrate CLI). Humans: see CONTRIBUTING.md and ARCHITECTURE.md.
Build, test, and check
pnpm install
pnpm build # bundles grammars, then tsup
pnpm test # vitest
pnpm lint # eslint
pnpm typecheck # tsc --noEmit
pnpm dev <args> # run the CLI from source (tsx src/cli.ts)
A change is not done until pnpm test, pnpm lint, and pnpm typecheck all
pass.
Prefer vg over grepping
This repo is the code-graph tool — use it on itself. Instead of blindly grepping or reading whole files, build the graph and query it. It is faster and costs fewer tokens:
pnpm dev build .
pnpm dev ask "where is the worker pool created?"
pnpm dev show <symbol> # a symbol and its relations
pnpm dev impact <symbol> # what breaks if I change this
pnpm dev path A B # how two symbols connect
pnpm dev tree # module/area overview
Use grep only for things the graph does not model (config files, comments, strings).
Determinism rules (non-negotiable)
Identical input must produce identical output (graph.json, reports). When
editing anything that reaches serialized output:
- Derive IDs from content (
src/engine/hash.ts,ids.ts) — never from order, memory address, or time. - Sort before serializing; never emit raw
Map/Setiteration order. - No
Date.now(), no unseeded randomness (usesrc/engine/rng.ts), no reliance on filesystem enumeration order. - Add or update a test asserting stable output for any graph-affecting change.
Public-only rule
This is the open, Apache-2.0 distribution. Do not add references to private
or internal Vibgrate systems, packages, or proprietary concepts (e.g. any
internal specification format, an extract command, or commercial-only
internals). Keep all code, docs, and comments scoped to the public CLI:
deterministic code graph, local MCP server, drift reporting, version-correct
library docs, and agent install. If a feature would require a non-public package,
stop and leave it out.
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 · 88 lines · 953 tokens per session scan A 196d6f727f39
cli AGENTS.md is an instructions file published in the GitHub repository vibgrate/cli (3 stars, last pushed 3d ago), licensed Apache-2.0. It adds 953 tokens to every session, about $0.0048 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
repo-cartographer CLAUDE.md
Claude Code instructions for builditwithgk/repo-cartographer, covering claude.md — repo-cartographer, what this is, core design principle (do not violate), tech and scope (mvp — must stay small, 2 day build).
Understand-Anything CLAUDE.md
Claude Code instructions for Egonex-AI/Understand-Anything, covering understand anything, project overview, prerequisites, architecture and dashboard.
nuxt-supabase-starter CLAUDE.md
Instructions for YuDefine/nuxt-supabase-starter, covering claude.md, language, project structure and meta vs template 邊界.
aspens AGENTS.md
Instructions for aspenkit/aspens, covering aspens, skills, commands, release and conventions.
kaiord AGENTS.md
Instructions for pablo-albaladejo/kaiord, covering kaiord, purpose, key files, subdirectories and for ai agents.
Sync CLAUDE.md
Instructions for stroland02/Sync, covering sync, the governing principle, environment — these will break your session, non-negotiables and how we work.