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/0xagentkitchen/kitchenloop/agents-mdgit clone --depth 1 https://github.com/0xagentkitchen/kitchenloopWhat 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 | $0.01113 | $0.01113 |
| Opus 5 | $0.00557 | $0.00557 |
| Sonnet 5 | $0.00223 | $0.00223 |
| Haiku 4.5 | $0.00111 | $0.00111 |
Grade A, and why
kitchenloop 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.
How it starts
The opening of the file, as written. The whole thing — 108 lines — stays where its author put it; the contents beside it link to each section on GitHub.
KitchenLoop
Autonomous improvement loop for any codebase, powered by Claude Code. Runs six phases in sequence: Backlog → Ideate → Triage → Execute → Polish → Regress. Each iteration picks work, implements it, tests it adversarially, and opens a PR — all without human intervention.
Key Concepts
- Spec surface — the matrix of features x platforms x actions your product claims to support
- Unbeatable tests — end-to-end verification against ground truth that the code author cannot fake
- UAT gate — post-implementation adversarial testing by a fresh agent with zero context
- Drain mode — auto-triggered when PR backpressure exceeds a configured threshold
- Discussion Manager — multi-AI deliberation with anti-sycophancy safeguards
Architecture
scripts/kitchenloop/
kitchenloop.sh # Main orchestrator (~2000 lines bash). Runs phases via Claude Code subprocess.
prompts/ # Phase prompts: ideate.md, execute.md, triage.md, polish.md, regress.md, backlog.md
# Uses template variables: {{REPO_ROOT}}, {{WORKTREE}}, {{BRANCH}}, etc.
lib/ # Internal helpers for the orchestrator
lib/
config.sh # yq-based config loader. Reads kitchenloop.yaml.
paths.sh # Artifact path resolver. All paths go through here.
tickets.sh # Ticket adapter (GitHub Issues / Linear).
scripts/ai-discussion/
discuss.py # Multi-AI discussion orchestrator (Python 3.10+). Anti-sycophancy safeguards.
scripts/pr-manager/ # Automated PR lifecycle: create, label, merge, drain.
skills/ # Source skill definitions. Copied to .claude/skills/ by the init script.
# Skills in .claude/skills/ are auto-discovered by Claude Code. Each has a SKILL.md.
# The prompts in scripts/kitchenloop/prompts/ are lightweight versions;
# the full skills have more detail.
templates/ # Template files for new project scaffolding.
examples/ # Example configs (Python CLI, Web API).
docs/howto.md # Operator manual for running KitchenLoop.
kitchenloop.example.yaml # Reference configuration with all available options.
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 · 108 lines · 1,113 tokens per session scan A 32496e0ba157
kitchenloop AGENTS.md is an instructions file published in the GitHub repository 0xagentkitchen/kitchenloop (23 stars, last pushed 5mo ago), licensed MIT. It adds 1,113 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.
Other instructions, from other repositories
codex AGENTS.md
AGENTS.md instructions for openai/codex, covering rust/codex-rs, the codex-core crate, code review rules, crate api surface and model visible context.
buildNext
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).
next.js AGENTS.md
Instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.
vscode oss-third-party-notices.instructions.md
Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, applying the notice (cutover) and fallback chain (never fail the build).
spec-kit 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.
langchain 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.