guild: Instructions file for Claude Code

AGENTS.md

guild AGENTS.md is an instructions file for Claude Code, Codex, OpenCode from lookatitude/guild. It costs 8,619 tokens per session, scanned A, original, MIT.

Repository instructions for Guild, a plugin that provides planning, coding, review, knowledge-management, and self-improvement workflows for coding agents. They describe its agents, skills, project files, and documentation layout.

In plain words
What is it for?
Use them when developing Guild skills, agents, plugin manifests, project state, review workflows, or its documentation.
Why use it?
They give agents the structure and workflow rules needed to navigate Guild and keep its shared instructions consistent across different hosts.

Instructions file for Claude CodeCodexOpenCode

Written for Claude Code and Codex and OpenCode: SessionStart hook event, but also the file is AGENTS.md. Also seen: mentions .claude-plugin; mentions CLAUDE.md; mentions Claude Code.

This is lookatitude/guild's own configuration. It tells Claude Code, Codex and OpenCode how to work on guild itself, so it is not a mod to install elsewhere. Copy it as a starting point and replace the rules that are about this project. Everything guild configures →

Reuse

Borrowing it

Nothing to install: this file belongs to lookatitude/guild. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file
curl -O https://raw.githubusercontent.com/lookatitude/guild/main/AGENTS.md
Clone the repo
git clone --depth 1 https://github.com/lookatitude/guild

Made for: Claude Code, Codex, OpenCode.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for guild AGENTS.md

README.md
[![agentmods](https://agentmods.dev/badge/instructions/lookatitude/guild/agents-md/github.svg)](https://agentmods.dev/instructions/lookatitude/guild/agents-md)
Your own site
<a href="https://agentmods.dev/instructions/lookatitude/guild/agents-md"><img src="https://agentmods.dev/badge/instructions/lookatitude/guild/agents-md/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for guild AGENTS.md

Your own site · 80×15
<a href="https://agentmods.dev/instructions/lookatitude/guild/agents-md"><img src="https://agentmods.dev/badge/instructions/lookatitude/guild/agents-md.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 8,619 This file is loaded in full into every session.
When invoked 8,619 The same file — it is already loaded in full.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
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.

ModelPer sessionOnce invoked
Fable 5.1 $0.08619 $0.08619
Opus 5.5 $0.03448 $0.03448
Sonnet 5.5 $0.01724 $0.01724
Haiku 4.5 $0.00862 $0.00862

Measured 20d ago against content hash cb6b9f3b4b57, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

guild 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 20d 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 · 434 lines

How it starts

The opening of the file, as written. The whole thing — 434 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Guild — repo orientation

Guild is a cross-host plugin that ships 4 machinery agents (advisor, context-manager, developer, team-lead), 15 domain specialist type templates (minted into a project's .guild/agents/ on demand by team composition), and 76 skills — 17 of them indexed — across a brainstorm-plan-execute-review-verify-reflect spine, a categorized wiki with decision capture, and a self-evolution loop with shadow-mode gating.

The runtime shape (bare /guild = T0, three tiers, five class graphs, one review gate, one promotion law, two context sizes) is the Layout laws list below; README carries the user-facing summary.

This AGENTS.md file is the canonical host-neutral instruction file. Claude Code loads CLAUDE.md, but CLAUDE.md must only import this file; all durable directions belong here so Codex, Pi, Antigravity, and AGENTS.md-consuming hosts read the same guidance.

For full architecture and design documentation see https://guildstack.dev/docs.

Where things live

  • .claude-plugin/plugin.json, .claude-plugin/marketplace.json — plugin + marketplace manifests.
  • skills/ — the skill tree. .claude-plugin/plugin.json indexes exactly the 17 assembler directories (KTD59); everything else ships off-index. An assembler is a three-stage folder: SKILL.md (frontmatter always, body on match) + references/ (L3 chapters, on demand) + scripts/ (compiled, never prompt text). The 58 specialist starter recipes live under skills/specialists/ as copy-on-mint feedstock (KTD13/KTD20) and the dashboard launcher under skills/playbooks/dashboard/; neither is indexed. The former core/ and fallback/ tiers are gone: principles folded into using-guild, and tdd / systematic-debug / worktrees / finish-branch are now chapters of the assembler that invokes them.
  • agents/*.md — the 4 machinery agents (advisor, context-manager, developer, team-lead), the only host-registered agents the plugin ships. Populated and authored.
  • templates/specialists/*.md — the 15 domain specialist type templates (guild.specialist_template.v1; architect … sales, incl. doc-writer), read-only feedstock minted into a project's .guild/agents/ by roster-resolve.ts mint during team composition.
  • commands/*.md — the v2 flat-token command surface (/guild:<verb>; the : plugin namespace stays — Claude Code requires it — v2 only drops the redundant guild- prefix; sub-verbs are positional arguments, never separate files or namespaces).
  • hooks/hooks.json — native Claude Code hooks; the .ts sources compile to the committed esbuild bundles under hooks/dist/ (+ hooks/agent-team/dist/) the host actually runs (cd hooks && npm run build after any hook edit — source edits are a no-op until rebuilt). Other hosts consume equivalent behavior through host adapters and graceful fallback.
  • src/domains/<id>/ — the twelve domains (KTD1/KTD36): kernel, state, security, config, lifecycle, knowledge, teams, dispatch, review, telemetry, evolve, distribution. Each has ONE public index.ts — nothing outside the domain may reach a sibling file — with internals and colocated tests beside it. The KTD36 fold map is data, at src/domains/distribution/domain-fold.ts, and is recorded in the coverage file.
  • src/adapters/ — host family maps and the remaining host-runtime projection (KTD4). Not a thirteenth domain: adapters may read a domain index, a domain must not import adapters.
  • src/runtime/mcp/ — the source of the two optional MCP servers; it compiles into the one committed runtime/guild-mcp.js with two D-MCP ids (wiki | trace, KTD3). src/surfaces/ holds only the five class graphs (graphs/*.yaml) and the prompt dialects (prompts/).
  • There is no src/modules/ tree: each former module's manifests and code live beside the domain it folded into.
  • Host packages are a PROJECTION of the live surface (KTD28): build-host-packages copies straight from commands/, skills/, agents/, hooks/, scripts/. There are no resources/ byte mirrors and no mirror-sync step.
  • scripts/ — author-plane CLIs (compile, lint, evolve, telemetry); run under bun, shipped to users as compiled Node under runtime/scripts/.
  • mcp-servers/ — tests and fixtures for the two MCP servers; their source is src/runtime/mcp/.
  • dist/ — committed per-host packages (claude-code, codex, pi, …) rendered by scripts/build-host-packages.ts; regenerated, never hand-edited.
  • tests/ — cross-cutting evals, fixtures, and the Verification Contract (tests/verification-contract/index.json). Every test runs under bun test --isolate; there is no Jest and no layout-laws baseline.
  • templates/{skills,agents,products}/ — authoring + product scaffolds (templates/specialists/ is the 15 specialist type templates, described above).
  • docs/ — RETIRED as a docs set: holds only a static redirect page to https://guildstack.dev/docs plus the logo asset. The public docs are generated from the umbrella's docs/v2/ + the website; reference knowledge lives in .guild/wiki/. Apart from README.md, CONTRIBUTING.md, and CHANGELOG.md, no documentation lives in this repo.

Read the full file on GitHub · 434 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.

  1. 20d ago Changed · +130 lines · +4,119 tokens per session cb6b9f3b4b57
  2. 2mo ago Changed · -10 lines · -319 tokens per session f3af8b66367c
  3. 2mo ago First seen · 314 lines · 4,819 tokens per session scan A 6bbd92b6b02b

Subscribe to this mod's changes

guild AGENTS.md is an instructions file published in the GitHub repository lookatitude/guild (8 stars, last pushed yesterday), licensed MIT. It adds 8,619 tokens to every session, about $0.0345 per session on Opus 5.5. 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.

Related

Other instructions, from other repositories

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 · 9,719 tokens

next.js AGENTS.md

AGENTS.md instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.

vercel/next.js · 7,461 tokens

vscode oss-third-party-notices.instructions.md

Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, cache eligibility and rollout and applying the notice (cutover).

microsoft/vscode · 5,516 tokens

langchain AGENTS.md

AGENTS.md instructions for langchain-ai/langchain, covering global development guidelines for the langchain monorepo, corridor security analysis, project architecture and context, monorepo structure and development tools & commands.

langchain-ai/langchain · 4,680 tokens

vscode copilot-instructions.md

Copilot instructions for microsoft/vscode, covering vs code copilot instructions, project overview, root folders, core architecture (src/ folder) and built-in extensions (extensions/ folder).

microsoft/vscode · 3,048 tokens

generative-ai-for-beginners AGENTS.md

AGENTS.md instructions for microsoft/generative-ai-for-beginners, covering agents.md, project overview, setup commands, initial repository setup and clone the repository.

microsoft/generative-ai-for-beginners · 2,850 tokens