What is forge AGENTS.md?

Repository instructions for the Forge project, covering its allowed branch, source-of-truth rules, release checks, and OpenClaw plugin handling.

Use them when working in projects/forge, preparing Forge releases, or changing the OpenClaw plugin.

They prevent changes from being made on the wrong branch or based on an outdated mirrored copy.

forge AGENTS.md — Instructions file by albertbuchard

Install

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.

agentmods

npx agentmods add instructions/albertbuchard/forge/agents-md

Clone the repo

git clone --depth 1 https://github.com/albertbuchard/forge

Made for: Codex, OpenCode.

Per session 1,118 This file is loaded in full into every session.

When invoked 1,118 The same file — it is already loaded in full.

Security scan A 0 findings. Scan, not verified.

Origin original No closer match found in the catalogue.

Token cost

What 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	$0.01118	$0.01118
Opus 5	$0.00559	$0.00559
Sonnet 5	$0.00224	$0.00224
Haiku 4.5	$0.00112	$0.00112

Measured 2d ago against content hash 3bf39b777d56, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

forge 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.

AGENTS.md · 53 lines

How it starts

The opening of the file, as written. The whole thing — 53 lines — stays where its author put it; the contents beside it link to each section on GitHub.

AGENTS.md — Forge

Scope

This file governs the projects/forge subtree and supplements the root AGENTS.md.

Branch rule

All Forge work must happen on main unless Albert explicitly asks for a different branch. Do not create, switch to, or keep working on feature branches by default. Before any Forge commit, verify git branch --show-current returns main.

Parent monorepo mirror rule

The Forge repository is the source of truth for files under projects/forge. The parent monorepo may mirror Forge files, but parent merge resolution is not authoritative for Forge source.

Do not let a parent-monorepo checkout, merge, or conflict resolution downgrade Forge files.
If parent and Forge disagree, inspect Forge main first and restore from the Forge repository's HEAD.
Before staging Forge release metadata, run scripts/release/audit-release-guard.sh.
The local Forge pre-commit hook also runs the release guard. It fails if the working or staged iOS marketing version is older than the latest ios-testflight-v* tag.

OpenClaw plugin rule

When Forge work involves the OpenClaw plugin, treat the repo-local plugin folder as the default install source.

Before changing, reinstalling, restarting, or debugging Forge, OpenClaw, Hermes, or the Forge Codex adapter, verify the effective data root from explicit config and live runtime evidence. Do not infer the database from generic defaults.
Check both config and runtime evidence when storage is in question: OpenClaw plugin config, Hermes Forge config, FORGE_DATA_ROOT, and a live process/file-handle check such as lsof -p <forge-pid> | rg 'forge.sqlite|forge.json'.
Treat fallback locations such as ~/.forge, plugin extension folders, package-local data/, and adapter-local data folders as candidates only until config or runtime evidence proves they are active.
Before any data-root repair, migration, restore, or merge, create a timestamped backup of every candidate Forge database and config file. Do not merge side databases into the selected canonical database unless an ID/content-level audit proves the side database contains relevant user data that is missing from that canonical database.
Use the documented local-dev flow from plugins/openclaw/README.md.
Prefer openclaw plugins install ./projects/forge/plugins/openclaw, then openclaw plugins enable forge-openclaw-plugin, then repair plugins.allow if needed, then openclaw gateway restart, then openclaw forge health.
Do not switch to the published npm package or another global install path unless Albert explicitly asks for that.
Do not change the Forge plugin dataRoot in ~/.openclaw/openclaw.json unless Albert explicitly asks for a data-path change.
When verifying state, confirm openclaw plugins info forge-openclaw-plugin reports the repo-local Forge plugin as the source path.
If Forge backend routes, the OpenClaw plugin package, onboarding contract, or Forge skill files changed, rebuild the plugin before verification so the running agent surface matches the source tree.
After such changes, do not stop at static inspection. Rebuild, restart the OpenClaw gateway, then re-check openclaw plugins info forge-openclaw-plugin and openclaw forge health.

Read the full file on GitHub · 53 lines

Changes

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 · 53 lines · 1,118 tokens per session scan A 3bf39b777d56

Subscribe to this mod's changes

forge AGENTS.md is an instructions file published in the GitHub repository albertbuchard/forge (23 stars, last pushed 5d ago), licensed Apache-2.0. It adds 1,118 tokens to every session, about $0.0056 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.

Git and workflow

spec-kit AGENTS.md

AGENTS.md instructions for github/spec-kit, covering agents.md, about spec kit and specify, quickstart — add a new integration in 5 steps, integration architecture and integrationmanifest — file tracking.

github/spec-kit · 7,104 tokens

vscode buildNext.instructions.md

Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).

microsoft/vscode · 6,785 tokens

forge AGENTS.md