What is ai-crew CLAUDE.md?

A set of project instructions for ai-crew, a Claude Code plugin that takes React Native or TypeScript backend tasks through planning, implementation, review, and delivery with one human checkpoint.

Use it when developing or maintaining ai-crew, especially when working with its vendored skills, agents, references, hooks, commands, and model-routing rules.

It defines the project’s fixed architecture, source-of-truth files, model assignments, and rules for changing the plugin safely.

How Goktug/ai-crew configures Claude Code

Reuse

Borrowing it

Nothing to install: this file belongs to Goktug/ai-crew. 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/Goktug/ai-crew/main/CLAUDE.md

Clone the repo

git clone --depth 1 https://github.com/Goktug/ai-crew

Made for: Claude Code.

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.

README.md

[![agentmods](https://agentmods.dev/badge/instructions/goktug/ai-crew/claude-md.svg)](https://agentmods.dev/instructions/goktug/ai-crew/claude-md)

Your own site

<a href="https://agentmods.dev/instructions/goktug/ai-crew/claude-md"><img src="https://agentmods.dev/badge/instructions/goktug/ai-crew/claude-md.svg" alt="Measured on agentmods" height="20"></a>

Per session 1,108 This file is loaded in full into every session.

When invoked 1,108 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.

Model	Per session	Once invoked
Fable 5.1	$0.01108	$0.01108
Opus 5	$0.00554	$0.00554
Sonnet 5	$0.00222	$0.00222
Haiku 4.5	$0.00111	$0.00111

Measured 7d ago against content hash 30898ddf21cd, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-07, from the pricing page.

Security

Grade A, and why

ai-crew 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 7d 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.

CLAUDE.md · 72 lines

How it starts

The opening of the file, as written. The whole thing — 72 lines — stays where its author put it; the contents beside it link to each section on GitHub.

ai-crew

Production-grade Claude Code plugin: takes any React Native or TypeScript backend task from idea to PR with one human checkpoint.

Design (locked)

All eight architectural decisions are locked. Do not propose changes to any of: model routing, dispatch tiers, artifact set, vendor strategy, or lifecycle phases without explicit approval.

Source of truth

reference-projects/agent-skills/ — base. Will be vendored verbatim (one-time copy) into skills/, agents/, references/, hooks/, and .claude/commands/ at the plugin root. Never edit vendored files.

Model routing

Layer	Model	Role
`team-lead` (main session)	Fable 5	Coordinator — plans big, executes small. Orchestrator + architect + planner + verifier + reviewer + shipper. Heavy thinking inline; never reads token-heavy raw material a dispatch can read.
`sonnet-developer` / `opus-developer` / `fable-developer` subagents	Sonnet / Opus / Fable 5	One task per dispatch, routed by the plan's `complexity` flag (`simple` / `complex` / `frontier`). `fable-developer` is rare by design — no rate arbitrage, only context isolation and depth. No `Agent`/`Task` tool.
`web-researcher` subagent	Haiku	One focused web-research question per dispatch. No `Agent`/`Task` tool.

Architectural laws (non-negotiable)

Strict 1-level dispatch. Subagents never spawn subagents.
Reference-based dispatch with line-range citations. Subagent prompts contain task IDs, skill paths, and line-range citations into spec.md and plan.md — never embedded content and never whole-file pointers. spec.md carries a Section Index and plan.md carries a Task Index; every developer dispatch includes a Plan task range and at least one Spec refs range so the subagent reads only the slices it needs.
Three-file disk artifacts per run: spec.md, plan.md, progress.md. Nothing else unless materially needed.
Run state at ~/.claude/ai-crew/runs/<YYYY-MM-DD-slug>/.
Vendored agent-skills SKILL.md content is never edited. One-time copy, no upstream sync.
Plan quality > parallelism. Atomized DAG with per-task acceptance criteria + verification commands. Sonnet executes mechanically when the plan is right.
No reviewer subagents. Fable team-lead does five-axis review inline using agent-skills' four review skills (code-review-and-quality, security-and-hardening, code-simplification, performance-optimization).
Verify/review retry policy: max 3 fix loops before escalating to user.

Read the full file on GitHub · 72 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.

7d ago First seen · 72 lines · 1,108 tokens per session scan A 30898ddf21cd

Subscribe to this mod's changes

ai-crew CLAUDE.md is an instructions file published in the GitHub repository Goktug/ai-crew (6 stars, last pushed 1mo ago), licensed MIT. It adds 1,108 tokens to every session, about $0.0055 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 2026-08-31.

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.

openai/codex · 5,182 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

ai-crew: Instructions file for Claude Code