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/forgeailab/forge/claude-mdgit clone --depth 1 https://github.com/ForgeAILab/forgeWhat 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 | $0.02122 | $0.02122 |
| Opus 5 | $0.01061 | $0.01061 |
| Sonnet 5 | $0.00424 | $0.00424 |
| Haiku 4.5 | $0.00212 | $0.00212 |
Grade A, and why
forge 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 — 123 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.
The user-facing entry point is README.md. Deeper material is split across docs/:
- docs/architecture.md — crate graph, state machine, workflow engine, DB.
- docs/api.md — REST surface, pagination, MCP tools.
- docs/getting-started.md — install, configuration, end-to-end walkthrough.
- docs/cli.md —
forge-ctlreference.
When you change behavior, update the relevant docs/ file in the same change. Treat README as a launch landing page — keep it short and link out.
Project status: Public Beta (0.1.x)
Forge is in public beta. The local-first single-user product is shipping to real users, but we are still pre-1.0 and pre-stable.
What that means for changes you make:
- Breaking changes are allowed, but they must be visible: update CHANGELOG.md under the
Unreleasedsection with a### Breakingentry, and bump the workspace version when cutting a release. Silent breakage is no longer acceptable now that users exist. - No backward-compat shims, no
_v2suffixes, no deprecated aliases. When something changes, update all call sites and delete the old code. Express the break in the changelog, not in the source. - No feature flags for rollout safety. Ship the new behavior directly.
- Database migrations must preserve user data unless the user explicitly opts in to a destructive migration (e.g. a hard schema reset for the dev
./testdata dir). Add a new numbered migration; do not edit historical ones. - Public surfaces — REST endpoints, MCP tool names,
forge-ctlflag names, event payload shapes, JSONL log schema, on-disk file layout — are the ones to be most deliberate about. Internal Rust APIs and crate boundaries can still move freely. - When a spec/design changes, rewrite the affected code to match. Don't layer the new design on top of the old one.
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 · 123 lines · 2,122 tokens per session scan A 2733912aee2d
forge CLAUDE.md is an instructions file published in the GitHub repository ForgeAILab/forge (54 stars, last pushed 16d ago), licensed MIT. It adds 2,122 tokens to every session, about $0.0106 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
diri AGENTS.md
Instructions for cristicretu/diri, covering diri repository instructions, scope, completed remote architecture baseline, rust workspace map and remote architecture invariants.
ephemeral-sandbox CLAUDE.md
Instructions for Ephemeral-AI-Lab/ephemeral-sandbox, covering claude.md, project, engineering practice (required), build & test and sandbox tools.
aristo CLAUDE.md
Instructions for aretta-ai/aristo, covering claude.md — aristo working agreement, §1. commit size — small or medium only, §2. commit messages — semantic / conventional, §3. changelog.md — one line per commit, in the same commit and §4. test-first — no test, no claim of correctness.
ephemeral-sandbox AGENTS.md
Instructions for Ephemeral-AI-Lab/ephemeral-sandbox, covering agents.md and required rules.
TreeWork AGENTS.md
Instructions for Johnny-xuan/TreeWork, covering treework repository guidance, read before editing, repository boundaries and change discipline.
RepoGrammar AGENTS.md
AGENTS.md instructions for SioYooo/RepoGrammar, covering repository agent contract, authority, repository boundaries, change discipline and verification.