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/vouchdev/vouch/claude-mdgit clone --depth 1 https://github.com/vouchdev/vouchWhat 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 | $0.03221 | $0.03221 |
| Opus 5 | $0.01611 | $0.01611 |
| Sonnet 5 | $0.00644 | $0.00644 |
| Haiku 4.5 | $0.00322 | $0.00322 |
Grade A, and why
vouch 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 — 298 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md — orientation for Claude Code working on vouch
This file is read automatically by Claude Code when you open the vouch repository. It exists to make a fresh session productive without you having to re-onboard it every time.
If you're working with vouch from inside a different project (proposing
claims, approving them), read README.md instead. This
file is for working on vouch — fixing a bug, adding a feature,
shipping a release.
For any non-Claude-Code agent — start with AGENTS.md.
North star
Vouch is a knowledge base where every write goes through a review gate. That's the load-bearing invariant. Every other design choice — files on disk, append-only audit log, manifest-driven adapters, thin viewports over the storage layer — is downstream of "writes must be reviewed."
If a PR adds a parallel data path that bypasses proposals.approve(),
the PR is wrong. Push back. Find the right factoring.
Architecture (90 seconds)
┌──────────────┐
Claude Code ─MCP──▶ │ │
Cursor ─MCP──▶ │ server.py │
Codex ─MCP──▶ │ jsonl_ │ ─┐
CLI human ──────▶ │ server.py │ │
│ cli.py │ │
└──────┬───────┘ │
│ │
▼ │
┌──────────────┐ │
│ proposals.py │ │ review gate
│ lifecycle.py │ │ (kb.approve etc.)
└──────┬───────┘ │
│ │
▼ │
┌──────────────┐ │
│ storage.py │ ◀──┘
│ audit.py │
│ index_db.py │
└──────┬───────┘
│
▼
.vouch/ ─── filesystem (yaml + md + jsonl)
─── state.db (FTS5 + optional embeddings, derived)
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 · 298 lines · 3,221 tokens per session scan A 16e31fde9ab1
vouch CLAUDE.md is an instructions file published in the GitHub repository vouchdev/vouch (91 stars, last pushed 10d ago), licensed MIT. It adds 3,221 tokens to every session, about $0.0161 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
nullclaw AGENTS.md
Instructions for nullclaw/nullclaw, covering agents.md — nullclaw agent engineering protocol, 1) project snapshot (read first), 3) engineering principles (normative), 3.1 kiss and 3.2 yagni.
nullclaw CLAUDE.md
Instructions for nullclaw/nullclaw, covering claude.md, mandatory reference, build & test commands, requires exactly zig 0.16.0 (verify: zig version) and build flags.
openclaw AGENTS.md
AGENTS.md instructions for openclaw/openclaw, covering agents.md, start, repair doctrine, product doctrine and clawsweeper review policy.
openclaw copilot.instructions.md
Copilot instructions for openclaw/openclaw, covering openclaw codebase patterns, tech stack, anti-redundancy rules, source of truth locations and formatting utilities (src/infra/).
openclaw CLAUDE.md
Claude Code instructions for openclaw/openclaw, a project described as: Your own personal AI assistant. Any OS. Any Platform. The lobster way. 🦞.
FERAL-AI CLAUDE.md
Instructions for FERAL-AI/FERAL-AI, covering claude.md — feral / asos, what this is, layout, commands and rather than restating it.