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/drobins25/craft/claude-mdgit clone --depth 1 https://github.com/drobins25/craftWrote 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.
[](https://agentmods.dev/instructions/drobins25/craft/claude-md)<a href="https://agentmods.dev/instructions/drobins25/craft/claude-md"><img src="https://agentmods.dev/badge/instructions/drobins25/craft/claude-md.svg" alt="Measured on agentmods" height="20"></a>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 | $0.00734 | $0.00734 |
| Opus 5 | $0.00367 | $0.00367 |
| Sonnet 5 | $0.00147 | $0.00147 |
| Haiku 4.5 | $0.00073 | $0.00073 |
Grade A, and why
craft 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 5d 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 — 36 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Craft
Open-source Claude Code plugin: a creative-first development harness.
- README.md - what craft is and how to use it
- DESIGN.md - architecture, file layout, internals
- CONTRIBUTING.md - contribution workflow for craft itself
This file is the operating contract for Claude working in this repo. Routing and live state are injected on every prompt via the UserPromptSubmit hook (the v1|craft-orchestration-index block) - don't duplicate that here.
Safety rules
- Never use
/craft:story-implementin this repo. It invokes the implementer agent against the plugin's own files, which is self-modification mid-work and breaks the harness. Contributors: see CONTRIBUTING.md for the manual implementation procedure. - Walkthrough quick-fixes are the only exception to "always invoke implementer via Task." Findings with
complexity: quick-fixand afix_hintcan be implemented directly by the orchestrator. Scope: trivial CSS or attribute edits, 1-5 lines, single file. Story-fix findings still go through the implementer agent.
Conventions
- Commit messages describe what changed for a public audience. Use conventional prefixes:
feat:/fix:/chore:/refactor:/docs:/test:. Translate internal jargon to what the change DOES. Public craft terms (skills, agents, commands, hooks, cycles, stories, backlog, phases) are fine; internal mechanism names need explanation.- Bad:
fix: chain break when content-spark invokes via Skill tool - Good:
fix: prevent skill nesting from breaking control flow back to caller
- Bad:
- Use regular dashes (-) only, never em dashes. Em dashes are an AI-text tell.
- Bump
.claude-plugin/plugin.jsonversion once per feature or fix - one bump per logical change (a story, an adhoc fix), in that change's final commit. Never per chunk; intermediate chunk commits do not bump. CHANGELOG.mdentries are notable-only - features and user-visible fixes get a## <version> - <YYYY-MM-DD>heading with verb-first, user-facing bullets ("Added...", "Fixed...", "Changed...", "Improved..."). Internal changes (tests, refactors, doc drift, contributor tooling) bump the version with no entry, and no entry is ever about the repo itself (changelog edits, doc restructures, contributor tooling). Same audience rule as commit messages.check-doc-drift.shblocks a changelog that's ahead of plugin.json, and blocks pushing afeat:commit with no changelog change.- Write every bullet for a craft user, not for us. The test: no internal mechanism names (CDN hosts, metadata fields, gate names, file paths), lead with what the user gets, and say what it feels like to use ("icons now arrive in seconds instead of minutes"), not how it works. The
## 1.99.54entry is the exemplar - match its register.
- Write every bullet for a craft user, not for us. The test: no internal mechanism names (CDN hosts, metadata fields, gate names, file paths), lead with what the user gets, and say what it feels like to use ("icons now arrive in seconds instead of minutes"), not how it works. The
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.
- 5d ago First seen · 36 lines · 734 tokens per session scan A dc02eb798c06
craft CLAUDE.md is an instructions file published in the GitHub repository drobins25/craft (53 stars, last pushed 5d ago), licensed MIT. It adds 734 tokens to every session, about $0.0037 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
claude-code-tresor CLAUDE.md
Instructions for alirezarezvani/claude-code-tresor, covering claude.md - development guide, 🎯 repository purpose, 🏗️ architecture, 🛠️ common development commands and building & testing.
ai-methods CLAUDE.md
Claude Code instructions for pirona/ai-methods, covering claude.md — claude code operator template, 1. identity & operator context, 2. homelab infrastructure, hypervisor & compute and network.
ai-agents push-lock.instructions.md
Instructions for rjmurillo/ai-agents, covering one push lock path, must, must not, historical measurements and commit guard.
agent-rules child-process.instructions.md
Instructions for lirantal/agent-rules, covering system processes secure coding guidelines, your mission and spawning system processes.
bdb-dev-optimized-agent-skills CLAUDE.md
Claude Code instructions for hybridlabor-api/bdb-dev-optimized-agent-skills, covering bdb agent skills — global instructions, docs & pipeline, how many agents, safety gate — mechanically enforced, not advisory and non-negotiable.
Myco AGENTS.md
Instructions for BreetyGreen/Myco, covering agents.md — orientation for ai agents & contributors, what this project is, architecture (two layers), single source of truth: engine/agents.json and key files.