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/dwarvesf/dwarves-kit/claude-mdgit clone --depth 1 https://github.com/dwarvesf/dwarves-kitWrote 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/dwarvesf/dwarves-kit/claude-md)<a href="https://agentmods.dev/instructions/dwarvesf/dwarves-kit/claude-md"><img src="https://agentmods.dev/badge/instructions/dwarvesf/dwarves-kit/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 | $0.01202 | $0.01202 |
| Opus 5 | $0.00601 | $0.00601 |
| Sonnet 5 | $0.00240 | $0.00240 |
| Haiku 4.5 | $0.00120 | $0.00120 |
Grade A, and why
dwarves-kit 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 4d 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 — 71 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md
Project instructions for working on dwarves-kit itself (the kit's own repo). If you are looking for the template to copy into a NEW project that consumes the kit, see examples/hello-spec/CLAUDE.md.
Project
dwarves-kit is a Claude Code workflow toolkit: hooks, slash commands, subagents, and a skill. Spec-driven lifecycle (/kit:think → /kit:spec → /kit:execute → /kit:review → /kit:ship → /kit:retro) with a verification pipeline (worker → task-verifier → fix-agent retry, max 2). Distributed as a Claude Code plugin and as a bash installer. Maintainer: Han at Dwarves Foundation.
For component fit and data flow, read docs/architecture.md. For operator detail per command, how to drive the kit from natural language, and how to recover from hook misbehavior, read MANUAL.md. Design rejection rules live in docs/PHILOSOPHY.md and are load-bearing.
Tech stack
| Layer | Choice |
|---|---|
| Hook runtime | bash + jq |
| Distribution | Claude Code plugin (.claude-plugin/) + bash installer (install.sh + settings.json) |
| Tests | bash (tests/test-hooks.sh for behavior, tests/test-meta.sh for structural integrity) |
| CI | GitHub Actions, macOS + Ubuntu matrix |
The kit deliberately uses no compiled binaries and no Node/Python in hooks (carve-out: statusline may use Node per PHILOSOPHY). Every script must be readable in 30 seconds.
Commands (for dev on the kit)
bash tests/test-hooks.sh # hook behavior tests
bash tests/test-meta.sh # structural integrity tests
bash install.sh # install into ~/.claude/ (idempotent)
bash install.sh --uninstall # clean removal
DWARVES_KIT_DEBUG=1 ... # verbose hook logging on stderr
The kit ships its own slash commands (/kit:start ... /kit:retro ... /kit:kit-health). Working on the kit means dogfooding: write a spec for your change, dispatch the verification pipeline, ship through /kit:ship. The kit's commands are listed in MANUAL.md; do not duplicate the inventory here.
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.
- 4d ago First seen · 71 lines · 1,202 tokens per session scan A f5bae312ac07
dwarves-kit CLAUDE.md is an instructions file published in the GitHub repository dwarvesf/dwarves-kit (11 stars, last pushed yesterday), licensed MIT. It adds 1,202 tokens to every session, about $0.0060 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
gangsta AGENTS.md
Instructions for kucherenko/gangsta, covering contributing to gangsta, what this is, what will not be accepted, skill changes and the laws apply to contributors too.
airflow CLAUDE.md
Claude Code instructions for apache/airflow, a project described as: Apache Airflow - A platform to programmatically author, schedule, and monitor workflows.
gangsta GEMINI.md
Instructions for kucherenko/gangsta, a project described as: AI agentic skills framework for spec-driven development, built on the organizational model of mafia.
hashloom CLAUDE.md
Claude Code instructions for davet47/hashloom, covering working rules for hashloom, git is the user's, scope discipline, hashloom develops on hashloom and definition of done: >5x token reduction.
pedal
Unified skill for managing the entire PEDAL cycle. PEDAL is a workspace-level workflow modeled after the traditional PDCA cycle. Supports 'Plan → Engineering → Do → Analyze → Learn' workflow with automatic phase progression. Use proactively when user mentions PEDAL cycle, planning, engineering documents, gap analysis…
claude-code-plugins CLAUDE.md
Claude Code instructions for bengous/claude-code-plugins, a project described as: Opinionated plugins and workflow extensions for Claude Code.