Skip to content
Back to search

spellbook

MarketplaceWorks with Claude CodeGitHub

Cross-runtime skill library and Claude Code plugins: product coaching, development toolkits, and workflow skills

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today

Check it deeper on your machine

It has changed since its scan on 11 Nov 2026.

As a Claude Code plugin, install it and then ask Claude Code:

/plugin marketplace add keithmackay/mackayi/plugin install scanskillsafety@mackayiDo a deep scan of https://github.com/majiayu000/spellbook with scanskillsafety

Or from a terminal, clone it and run it:

git clone https://github.com/keithmackay/scanskillsafety-skill ~/.claude/skills/scanskillsafetynode ~/.claude/skills/scanskillsafety/cli.cjs --deep 'https://github.com/majiayu000/spellbook'

It checks the exact commit: every file, its packages against OSV (api.osv.dev), how it compares with this page's scan (findsafeskills.com receives the repository's address), and, if you agree, Claude reading every file with your own Claude subscription. It never runs the code.

Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/majiayu000--spellbook/badge.svg)](/listings/majiayu000--spellbook)

HTML

<a href="/listings/majiayu000--spellbook"><img src="/api/listings/majiayu000--spellbook/badge.svg" alt="findsafeskills safety rating" /></a>

Install

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

These steps run a script from the network that findsafeskills does not scan.

Start with one job-shaped workflow. The maintained `skills` CLI lets you choose the supported coding agents during installation and installs only these four skills:

npx skills add majiayu000/spellbook --skill frontend-design --skill app-ui-design --skill ui-design-system --skill figma-to-react

Use `npx skills add majiayu000/spellbook --list` to inspect the catalog before installing. See [Pick a Workflow](#pick-a-workflow) for four other focused starting points.

Advanced Cross-Runtime Installer

`install.sh` remains available when you want explicit Claude Code/Codex target paths or need to install the repository's Claude Code agents as well as skills.

# Install all skills and supported agents into both maintained runtimes
curl -fsSL https://raw.githubusercontent.com/majiayu000/spellbook/main/install.sh | bash -s -- --target all

# Or clone the repository and select skills explicitly
git clone https://github.com/majiayu000/spellbook.git
cd spellbook
./install.sh --target all --skills typescript-project,python-project,devops-excellence

Verify Installation

- Claude Code: type `/` to see your installed skills. - Codex: restart Codex so it reloads `~/.agents/skills`.

---

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

.claude-plugin/marketplace.json · Manifest · 30 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

This file has changed since it was scanned, so this is not exactly the text that was scanned, and no line is marked. The findings are listed with the safety scan above.

  1. {
  2. "$schema": "https://anthropic.com/claude-code/marketplace.schema.json",
  3. "name": "spellbook",
  4. "description": "Cross-runtime skill library and Claude Code plugins: product coaching, development toolkits, and workflow skills",
  5. "strict": true,
  6. "owner": {
  7. "name": "majiayu000",
  8. "url": "https://github.com/majiayu000"
  9. },
  10. "plugins": [
  11. {
  12. "name": "idea-coach",
  13. "description": "Idea coaching toolkit: an opinionated end-to-end product coach (idea -> PRD -> clickable HTML prototype) plus a multi-role idea group chat for blind-spot discovery",
  14. "author": {
  15. "name": "majiayu000"
  16. },
  17. "category": "product",
  18. "source": "./plugins/idea-coach"
  19. },
  20. {
  21. "name": "rust-dev",
  22. "description": "Rust development toolkit with best practices, code review, performance optimization, and async Rust patterns",
  23. "author": {
  24. "name": "majiayu000"
  25. },
  26. "category": "development",
  27. "source": "./plugins/rust-dev"
  28. }
  29. ]
  30. }
README.md · README · 493 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

This file has changed since it was scanned, so this is not exactly the text that was scanned, and no line is marked. The findings are listed with the safety scan above.

  1. <div align="center">
  2. <h1>Spellbook</h1>
  3. <p><strong>121 Cross-Runtime Skills | 7 Claude Code Agents | One Command Install</strong></p>
  4. <p>A cross-runtime skill library for Claude Code, Codex, and multi-agent workflows.</p>
  5. <p>
  6. <a href="https://github.com/majiayu000/spellbook/stargazers"><img src="https://img.shields.io/github/stars/majiayu000/spellbook?style=flat-square&logo=github" alt="Stars"></a>
  7. <a href="https://github.com/majiayu000/spellbook/blob/main/LICENSE"><img src="https://img.shields.io/github/license/majiayu000/spellbook?style=flat-square" alt="License"></a>
  8. <img src="https://img.shields.io/badge/skills-121-blue?style=flat-square" alt="Skills">
  9. <img src="https://img.shields.io/badge/agents-7-green?style=flat-square" alt="Agents">
  10. </p>
  11. <p>
  12. <a href="#quick-start">Quick Start</a> •
  13. <a href="#runtime-targets">Runtime Targets</a> •
  14. <a href="#pick-a-workflow">Pick a Workflow</a> •
  15. <a href="#skills">Skills</a> •
  16. <a href="#agents">Agents</a> •
  17. <a href="./CHANGELOG.md">Changelog</a> •
  18. <a href="#release-status">Release Status</a> •
  19. <a href="#contributing">Contributing</a> •
  20. <a href="./README_CN.md">中文</a>
  21. </p>
  22. </div>
  23. ---
  24. > **Rename notice:** Spellbook was formerly **Claude Arsenal**. Claude Code remains a first-class target; the new name reflects the broader roadmap for Claude Code, Codex, and cross-runtime agent skills. See the [migration note](./docs/migration-from-claude-arsenal.md) for details.
  25. ---
  26. ## Quick Start
  27. Start with one job-shaped workflow. The maintained `skills` CLI lets you choose
  28. the supported coding agents during installation and installs only these four
  29. skills:
  30. ```bash
  31. npx skills add majiayu000/spellbook --skill frontend-design --skill app-ui-design --skill ui-design-system --skill figma-to-react
  32. ```
  33. Use `npx skills add majiayu000/spellbook --list` to inspect the catalog before
  34. installing. See [Pick a Workflow](#pick-a-workflow) for four other focused
  35. starting points.
  36. ### Advanced Cross-Runtime Installer
  37. `install.sh` remains available when you want explicit Claude Code/Codex target
  38. paths or need to install the repository's Claude Code agents as well as skills.
  39. ```bash
  40. # Install all skills and supported agents into both maintained runtimes
  41. curl -fsSL https://raw.githubusercontent.com/majiayu000/spellbook/main/install.sh | bash -s -- --target all
  42. # Or clone the repository and select skills explicitly
  43. git clone https://github.com/majiayu000/spellbook.git
  44. cd spellbook
  45. ./install.sh --target all --skills typescript-project,python-project,devops-excellence
  46. ```
  47. ### Verify Installation
  48. - Claude Code: type `/` to see your installed skills.
  49. - Codex: restart Codex so it reloads `~/.agents/skills`.
  50. ---
  51. ## Runtime Targets
  52. Spellbook keeps the skill source in one place and installs it into the runtime you use.
  53. | Target | Installed To | Status |
  54. |--------|--------------|--------|
  55. | Claude Code | `~/.claude/skills` plus `~/.claude/agents` | Skills and agents supported |
  56. | Codex | `~/.agents/skills` | Skills supported; agents skipped |
  57. | All | Both Claude Code and Codex paths | Recommended for multi-tool users |
  58. Claude Code remains a first-class target and search entry. The project was formerly known as Claude Arsenal; the new Spellbook name reflects the broader goal: reusable skills that can travel across coding agents.
  59. Older Spellbook versions installed Codex skills under `~/.codex/skills`; reinstall with the current installer to use the documented Codex user-level skill path.
  60. ---
  61. ## Use `.agents/skills` in Claude Code
  62. Checked on October 7, 2026: Claude Code's [instruction documentation](https://code.claude.com/docs/en/memory#agentsmd) supports `AGENTS.md` directly from v2.1.277. By default it loads these files only when no `CLAUDE.md`, `.claude/CLAUDE.md`, or `CLAUDE.local.md` exists in the working directory or its ancestors. You can choose to load both in the Project instructions setting. If direct support is unavailable, put `@AGENTS.md` in a neighboring `CLAUDE.md`.
  63. Skills use separate discovery paths: the [documented locations](https://code.claude.com/docs/en/skills#choose-where-skills-load) include `.claude/skills` and `~/.claude/skills`, but not `.agents/skills`. For a repository that already keeps its skills in `.agents/skills`, run this from the repository root **only if `.claude/skills` does not already exist**, including as a symlink:
  64. ```bash
  65. mkdir -p .claude
  66. ln -s ../.agents/skills .claude/skills
  67. ```
  68. Both paths now expose the same directory. Changes affect both tools, so you cannot add a Claude-only skill there. If you already have `.claude/skills`, keep it and link only the shared skill instead (replace `my-skill` with an existing skill directory):
  69. ```bash
  70. ln -s ../../.agents/skills/my-skill .claude/skills/my-skill
  71. ```
  72. The official skills documentation supports individual skill-directory symlinks. On macOS, Claude Code 2.1.234 listed a test skill in its CLI initialization event with either link layout; without a link, the same command was unknown. Model execution could not be verified because that installation was not logged in. Restart Claude Code, check `/` for your skill, and invoke it to verify your own environment.
  73. On Windows, [directory symlinks use `mklink /D`](https://learn.microsoft.com/en-us/windows-server/administration/windows-commands/mklink), and creating symlinks can require Developer Mode or Administrator privileges. Git's `core.symlinks` setting also affects checkout. Windows loading was not tested here; teams can use separate target entries instead:
  74. ```bash
  75. # From a Spellbook checkout; installs selected skills at user scope
  76. ./install.sh --target all --skills typescript-project,python-project
  77. # Or use the skills CLI and select Claude Code and Codex when prompted
  78. npx skills add majiayu000/spellbook --skill typescript-project
  79. ```
  80. The installer creates entries in `~/.claude/skills` and `~/.agents/skills` pointing to its shared source checkout. It does not synchronize arbitrary project directories. These instructions describe local Claude Code CLI use; other session types can have different loading rules.
  81. ---
  82. ## Pick a Workflow
  83. Start with a small bundle that matches the job, then add more skills when the workflow sticks.
  84. | Workflow | Install | Good for |
  85. |---|---|---|
  86. | Frontend and UI | `npx skills add majiayu000/spellbook --skill frontend-design --skill app-ui-design --skill ui-design-system --skill figma-to-react` | Product UI, landing pages, design systems, Figma handoff |
  87. | Code quality | `npx skills add majiayu000/spellbook --skill codebase-audit --skill flowguard --skill systematic-debugging --skill review-gate` | Audits, guarded delivery, root-cause debugging, pre-landing review |
  88. | Ops and release | `npx skills add majiayu000/spellbook --skill release-engineering --skill server-security --skill clash-doctor --skill system-doctor` | Release planning, server hardening, and local or network diagnosis |
  89. | Product and docs | `npx skills add majiayu000/spellbook --skill product-discovery --skill prd-master --skill technical-spec --skill product-analytics` | Discovery, PRDs, technical specs, metrics plans |
  90. | Agent workflows | `npx skills add majiayu000/spellbook --skill codex-agent --skill multi-ai-research --skill flowguard --skill vibeguard` | Cross-review, multi-AI research, context handoff, anti-hallucination checks |
  91. High-signal individual skills to try first: `github-trending`, `harmonyos-app`, `app-ui-design`, `product-discovery`, `xiaohongshu`, `codebase-audit`, and `server-security`.
  92. See [Showcase](./docs/showcase.md) for copy-paste prompts and expected outputs.
  93. Use the [Spellbook Skill Browser](https://majiayu000.github.io/spellbook/skills.html)
  94. for curated first-party skills, or the
  95. [Claude Skills Registry](https://majiayu000.github.io/claude-skill-registry/)
  96. for broader community discovery.
  97. Release history lives in [Changelog](./CHANGELOG.md).
  98. ---
  99. ## Why Spellbook
  100. - **Cross-runtime install**: one source tree can install into Claude Code and Codex.
  101. - **Validated registry**: every installable skill is checked by `python3 scripts/validate_skills.py --check`.
  102. - **Progressive disclosure**: larger skills use `references/`, `templates/`, `scripts/`, and eval files instead of one giant prompt.
  103. - **Practical coverage**: engineering, operations, product, UI, content, and agent workflows live in one catalog.
  104. ---
  105. ## Skills
  106. > The generated full skill inventory lives in [Skill Registry](./docs/skill-registry.md).
  107. > Skill layout rules live in [Skill Format Policy](./docs/skill-format-policy.md).
  108. > Skill authoring quality rules live in [Skill Quality Playbook](./docs/skill-quality-playbook.md).
  109. ### Search the Registry
  110. ```bash
  111. # Free-text query (AND semantics across name, description, category, tags)
  112. python3 scripts/validate_skills.py search rust testing
  113. # Filter by tag
  114. python3 scripts/validate_skills.py search --tag agent
  115. # Restrict to a description language
  116. python3 scripts/validate_skills.py search --language zh deploy
  117. # Machine-readable output
  118. python3 scripts/validate_skills.py search --tag react --json
  119. ```
  120. The tag index lives in [`registry/tags.json`](./registry/tags.json) for tooling and dashboards. Curated overrides for skills the keyword heuristic cannot infer live in [`registry/tag_overrides.yml`](./registry/tag_overrides.yml).
  121. Audit non-blocking skill quality signals:
  122. ```bash
  123. python3 scripts/audit_skill_quality.py
  124. python3 scripts/audit_skill_quality.py skill-creator
  125. ```
  126. ### AI & Agent Workflow
  127. Skills for orchestrating, guarding, and maintaining AI agent workflows — the core of Spellbook's cross-runtime mission.
  128. | Skill | Description |
  129. |-------|-------------|
  130. | [`multi-model-orchestrator`](./skills/multi-model-orchestrator/) | Coordinate multi-agent tasks via a centralized handoff document |
  131. | [`flowguard`](./skills/flowguard/) | Guard long, ambiguous, or stateful agent tasks from drift |
  132. | [`skill-lifeguard`](./skills/skill-lifeguard/) | Add reliable-skill contracts, checkpoints, smoke hooks, and drift signals |
  133. | [`review-gate`](./skills/review-gate/) | Produce review packs and require human approval before landing agent changes |
  134. | [`skill-audit`](./skills/skill-audit/) | Audit, design, categorize, and measure agent skills |
  135. | [`skill-checkup`](./skills/skill-checkup/) | Find broken, drifted, competing and one-sided skills across Claude Code and Codex |
  136. | [`skill-ecosystem-doctor`](./skills/skill-ecosystem-doctor/) | Govern canonical sources, projections, retirement, quarantine, and cross-runtime verification |
  137. | [`threads`](./skills/threads/) | Codex-native subagents and parallel GitHub queue lanes |
  138. | [`codex-fluent`](./skills/codex-fluent/) | Codex session hygiene, archive strategy, and handoff discipline |
  139. | [`codex-retrospective`](./skills/codex-retrospective/) | Codex self-review of recent history to improve behavior |
  140. | [`model-inference-optimize`](./skills/model-inference-optimize/) | Profile and optimize model inference with quality and cost checks, using anonymized cases |
  141. | [`brainstorming`](./skills/brainstorming.SKILL.md) | Socratic dialogue for design refinement and architecture exploration |
  142. See [`docs/agent-reliability-trio.md`](./docs/agent-reliability-trio.md) for the Reliable Skill + Context Engineering + Review Gate workflow.
  143. ### Development Architecture
  144. Build production-ready projects with language-specific best practices.
  145. | Skill | Language | Key Features |
  146. |-------|----------|--------------|
  147. | [`typescript-project`](./skills/typescript-project/) | TypeScript | ESM, Zod, Biome, Clean Architecture |
  148. | [`python-project`](./skills/python-project/) | Python | uv, Pydantic, Ruff, FastAPI |
  149. | [`rust-project`](./skills/rust-project/) | Rust | Cargo workspace, error handling, async |
  150. | [`golang-web`](./skills/golang-web/) | Go | Chi/Echo, sqlc, structured logging |
  151. | [`zig-project`](./skills/zig-project/) | Zig | Build system, memory management |
  152. | [`architecture-foundation`](./skills/architecture-foundation/) | Cross-language | Runtime, state ownership, adapters, and convergence specs |
  153. | [`elegant-architecture`](./skills/elegant-architecture/) | Cross-language | Clean architecture with strict 200-line file limits |
  154. ### Product Lifecycle
  155. End-to-end product development from discovery to deployment.
  156. | Skill | Phase | What You Get |
  157. |-------|-------|--------------|
  158. | [`product-discovery`](./skills/product-discovery/) | Discovery | JTBD, user interviews, market research |
  159. | [`prd-master`](./skills/prd-master/) | Definition | PRD writing, user stories, RICE prioritization |
  160. | [`technical-spec`](./skills/technical-spec/) | Design | Design docs, ADR, C4 diagrams |
  161. | [`product-analytics`](./skills/product-analytics/) | Growth | Event tracking, A/B testing, AARRR |
  162. | [`devops-excellence`](./skills/devops-excellence/) | Deployment | CI/CD, Docker, Kubernetes, GitOps |
  163. | [`observability-sre`](./skills/observability-sre/) | Operations | Monitoring, logging, tracing, SLO/SLI |
  164. | [`product-manager-toolkit`](./skills/product-manager-toolkit/) | Definition | RICE, customer interviews, PRD templates, discovery frameworks |
  165. ### API & Backend
  166. | Skill | Description |
  167. |-------|-------------|
  168. | [`api-design`](./skills/api-design/) | REST/GraphQL/gRPC patterns, OpenAPI 3.2 |
  169. | [`auth-security`](./skills/auth-security/) | OAuth 2.1, JWT, security best practices |
  170. | [`database-patterns`](./skills/database-patterns/) | PostgreSQL, Redis, migrations, optimization |
  171. | [`codebase-audit`](./skills/codebase-audit/) | Deep adaptive repository audit with severity-ranked findings and repair roadmap |
  172. | [`structured-logging-lite`](./skills/structured-logging-lite/) | Centralized logging, field standards, and distributed tracing |
  173. ### Development Practices
  174. | Skill | Description | Origin |
  175. |-------|-------------|--------|
  176. | [`contributor`](./skills/contributor/) | End-to-end open source contribution workflow from issue discovery to PR submission | Custom |
  177. | [`repo-agent-context-audit`](./skills/repo-agent-context-audit/) | Audit and scaffold repo agent context across AGENTS, skills, and specs | Custom |
  178. | [`skill-creator`](./skills/skill-creator/) | Create, improve, and benchmark reusable skills | Custom |
  179. | [`humanizer`](./skills/humanizer/) | Remove obvious AI writing patterns from user-facing text | External guide + custom adaptation |
  180. ### Delivery Workflow
  181. Disciplined end-to-end delivery: testing, commits, health checks, and contribution flow.
  182. | Skill | Description |
  183. |-------|-------------|
  184. | [`agent-qa-testing`](./skills/agent-qa-testing/) | Web/mobile regression runs and triage with advanced memidex memory and self-improvement |
  185. | [`app-user-story-qa`](./skills/app-user-story-qa/) | End-to-end app feature inventory, canonical tracker, user-story testing, fixes, and retest loop |
  186. | [`test-driven-development`](./skills/test-driven-development.SKILL.md) | Enforce RED-GREEN-REFACTOR TDD discipline |
  187. | [`comprehensive-testing`](./skills/comprehensive-testing/) | Test pyramid, unit/integration/E2E/property testing, framework best practices |
  188. | [`git-commit-smart`](./skills/git-commit-smart/) | Generate meaningful conventional commit messages from diff |
  189. | [`push-all`](./skills/push-all.SKILL.md) | Stage, commit, and push all changes after safety checks |
  190. | [`project-health-auditor`](./skills/project-health-auditor/) | Codebase health, tech debt, dependency, and project risk analysis |
  191. | [`loop-close`](./skills/loop-close/) | Close the loop on a feature, study or experiment: name the decision it supports and the quit condition up front, then end with adopt / reject / needs-evidence and archive it |
  192. | [`contribution-architect`](./skills/contribution-architect/) | Move from bug fixes to architectural improvements and debt discovery |
  193. ### Cross-Tool Interop
  194. Skills for using multiple coding agents and CLI tools together.
  195. | Skill | Description |
  196. |-------|-------------|
  197. | [`codex`](./skills/codex/) | Invoke Codex CLI sessions from another agent workflow |
  198. | [`codex-agent`](./skills/codex-agent/) | Optional second-opinion review, cross-verification, and alternatives through Codex CLI |
  199. | [`sol-luna-router`](./skills/sol-luna-router/) | Keep GPT-5.6 Sol as commander/reviewer while GPT-5.6 Luna performs bounded implementation |
  200. | [`ask-opencli`](./skills/ask-opencli/) | Ask Grok or Gemini through opencli and an existing browser session |
  201. | [`multi-ai-research`](./skills/multi-ai-research/) | Parallel research across multiple AI tools and internal agents |
  202. ### UI/UX & Design
  203. | Skill | Description |
  204. |-------|-------------|
  205. | [`app-ui-design`](./skills/app-ui-design/) | iOS/Android UI design, Material Design 3, HIG |
  206. | [`product-ux-expert`](./skills/product-ux-expert/) | UX evaluation, heuristics, accessibility |
  207. | [`frontend-design`](./skills/frontend-design/) | Web frontend design patterns |
  208. | [`illustrated-gallery`](./skills/illustrated-gallery/) | Illustrated website template with full-frame ASCII transitions |
  209. | [`ascii-backdrop`](./skills/ascii-backdrop/) | Animated ASCII backdrop from any subject: glyph, edge, shape-matched, braille, dither or tile modes, cursor glow, CRT finish, text decode and image dissolve |
  210. | [`code-built-3d`](./skills/code-built-3d/) | Build 3D scenes from code-made parts and take them apart: technical-illustration render, isometric camera moves, shell removal, exploded views with a parts list, endless chunked generation; demo is an endless apartment tower |
  211. | [`form-motion`](./skills/form-motion/) | Ten restrained, hand-following UI micro-interactions (morphing button, liquid tab indicator, elastic slider, command palette, drag reorder and more), each with its prompt, knobs and an automated pointer/keyboard acceptance script |
  212. | [`hud-interface`](./skills/hud-interface/) | Game-HUD style pages: ink-and-paper with one accent, slanted tags, giant outline numbers, ring text, tilted cards, a liquid-marble shader backdrop, and optional HTML-in-Canvas glass panels with a plain-DOM fallback |
  213. | [`ui-designer`](./skills/ui-designer/) | Extract design systems from UI screenshots and references |
  214. | [`ui-design-system`](./skills/ui-design-system/) | Design system toolkit and design-dev handoff support |
  215. | [`web-artifacts-builder`](./skills/web-artifacts-builder/) | Claude.ai HTML artifacts |
  216. | [`react-best-practices`](./skills/react-best-practices/) | React and Next.js performance patterns distilled from Vercel guidance |
  217. | [`react-hooks-best-practices`](./skills/react-hooks-best-practices/) | React hooks, effects, refs, and component design patterns |
  218. | [`blender-editable-3d`](./skills/blender-editable-3d/) | Editable Blender models from references or concepts, with part-level edits and verified exports |
  219. | [`blender-reference-to-3d`](./skills/blender-reference-to-3d/) | Reference analysis, Blender character modeling, color, multiview review, and editable delivery |
  220. | [`slides`](./skills/slides/) | Speech-friendly slide deck and background slide generation |
  221. | [`ui-ux-pro-max`](./skills/ui-ux-pro-max/) | Compact UI/UX tables for product patterns, landing pages, charts, and 9 stacks |
  222. | [`figma-to-code`](./skills/figma-to-code/) | Figma designs to production React/Next.js with TypeScript and Tailwind |
  223. | [`css-debug`](./skills/css-debug/) | Diagnose CSS/layout issues, Tailwind conflicts, z-index stacking |
  224. | [`playwright-automation`](./skills/playwright-automation/) | Browser automation and testing with Playwright |
  225. ### Tooling & Automation
  226. | Skill | Description |
  227. |-------|-------------|
  228. | [`web-asset-generator`](./skills/web-asset-generator/) | Favicons, app icons, OG images |
  229. | [`github-trending`](./skills/github-trending/) | GitHub trending analysis |
  230. | [`vibeguard`](./skills/vibeguard/) | Task contracts, finding scoring, and lightweight anti-hallucination reviews |
  231. | [`clash-doctor`](./skills/clash-doctor/) | Clash proxy & network diagnostics |
  232. | [`clash-routes`](./skills/clash-routes/) | Inspect active proxy routes for specific processes via Mihomo API |
  233. | [`optimize-network`](./skills/optimize-network/) | Safe local network speed, latency, DNS, Wi-Fi, and bufferbloat diagnostics with VPN/proxy guardrails |
  234. | [`disk-cleaner`](./skills/disk-cleaner/) | Scan and reclaim disk space with interactive cleanup guidance |
  235. | [`system-doctor`](./skills/system-doctor/) | Diagnose CPU, memory, and process-level system slowdowns |
  236. | [`codex-log-guard`](./skills/codex-log-guard/) | Diagnose and mitigate excessive Codex local SQLite diagnostic log writes |
  237. | [`server-security`](./skills/server-security/) | Audit and harden Linux server SSH, firewall, and exposed services |
  238. | [`cliproxy-newapi-stack`](./skills/cliproxy-newapi-stack/) | Add a loopback-first NewAPI metering layer to an independently verified CLIProxyAPI upstream |
  239. ### Operations & Deploy
  240. Deploy models and diagnose local and remote environments.
  241. | Skill | Description |
  242. |-------|-------------|
  243. | [`gemma4-local-deploy`](./skills/gemma4-local-deploy/) | Deploy Gemma 4 12B locally on Mac/Apple Silicon via llama.cpp or Ollama |
  244. | [`gpu-use`](./skills/gpu-use/) | Inspect remote server GPU usage (per-card VRAM, processes, containers) |
  245. | [`rustdesk-doctor`](./skills/rustdesk-doctor/) | Diagnose RustDesk connection issues |
  246. | [`vscode-doctor`](./skills/vscode-doctor/) | Diagnose slow or freezing VS Code-compatible editors |
  247. ### Content & Social Media
  248. | Skill | Description |
  249. |-------|-------------|
  250. | [`xiaohongshu`](./skills/xiaohongshu/) | Xiaohongshu content creation & publishing |
  251. | [`trip-planner`](./skills/trip-planner/) | Travel itinerary planning |
  252. | [`avoidance-coach`](./skills/avoidance-coach/) | Self-help coach for breaking avoidance loops (CBT/DBT/ACT, not therapy) |
  253. | [`code-film`](./skills/code-film/) | Make story-driven animated short films from pure code: a three.js HTML file whose picture and sound are functions of time, rendered offline to MP4, with the Clawd character included |
  254. | [`motion-stack`](./skills/motion-stack/) | Code-made motion films: pick the fewest of 20 open-source libraries, write a scene whose picture and sound are pure functions of time, render headless to MP4 with purity check, auto loudness (-14 LUFS) and a reproducibility manifest |
  255. | [`repo-film`](./skills/repo-film/) | Turn a repo's real commits over any time window into a short film: main line and branches grow as a flowchart with true forks and merges, optionally collapse into its you-are-here project map; one command to MP4 |
  256. | [`weekly`](./skills/weekly.SKILL.md) | Weekly report from Git, Claude Code, and Codex sessions |
  257. | [`xiaohongshu-netfeel-guardian`](./skills/xiaohongshu-netfeel-guardian/) | Remove translation-tone from Claude's Chinese content for native readability |
  258. ### Mobile & Cross-Platform
  259. | Skill | Description |
  260. |-------|-------------|
  261. | [`harmonyos-app`](./skills/harmonyos-app/) | HarmonyOS with ArkTS, ArkUI, Stage Model |
  262. ### Rust Specific
  263. | Skill | Description |
  264. |-------|-------------|
  265. | [`rust-best-practices`](./skills/rust-best-practices/) | Microsoft Rust guidelines, error handling |
  266. ---
  267. ## Agents
  268. Specialized agents for complex tasks.
  269. | Agent | Expertise | Use Case |
  270. |-------|-----------|----------|
  271. | [`tech-lead-orchestrator`](./agents/tech-lead-orchestrator.md) | Coordination | Multi-step tasks, delegation |
  272. | [`code-archaeologist`](./agents/code-archaeologist.md) | Exploration | Legacy codebase documentation |
  273. | [`backend-typescript-architect`](./agents/backend-typescript-architect.md) | Architecture | Bun/Node.js, API design |
  274. | [`senior-code-reviewer`](./agents/senior-code-reviewer.md) | Review | Security, performance, architecture |
  275. | [`kubernetes-specialist`](./agents/kubernetes-specialist.md) | Infrastructure | K8s, Helm, GitOps |
  276. | [`security-auditor`](./agents/security-auditor.md) | Security | OWASP Top 10, SAST |
  277. | [`opensource-contributor`](./agents/opensource-contributor.md) | Contribution | Open source workflow |
  278. ---
  279. ## Plugins
  280. Spellbook is also a Claude Code plugin marketplace. Install the repo as a
  281. marketplace, then install plugins from it:
  282. ```
  283. /plugin marketplace add majiayu000/spellbook
  284. /plugin install idea-coach
  285. /plugin install rust-dev
  286. ```
  287. | Plugin | Description |
  288. |--------|-------------|
  289. | [`idea-coach`](./plugins/idea-coach/) | Opinionated product coach (idea -> PRD -> clickable HTML prototype) + multi-role idea group chat; plugin commands are `/idea-coach:idea` and `/idea-coach:idea-team` |
  290. | [`rust-dev`](./plugins/rust-dev/) | Rust best practices, code review, performance, and async patterns |
  291. Plugin skills are packaged copies of catalog skills; the catalog (installed by
  292. `install.sh`) remains the cross-runtime source of truth.
  293. ---
  294. ## Skill Design Philosophy
  295. Every skill in Spellbook follows these principles:
  296. 1. **Hard Rules** - Mandatory constraints with `FORBIDDEN` / `REQUIRED` markers
  297. 2. **Practical Examples** - Real code, not just theory
  298. 3. **Verification Checklists** - Actionable validation steps
  299. 4. **Battle-Tested** - Used in production environments
  300. ---
  301. ## Documentation
  302. | Document | Description |
  303. |----------|-------------|
  304. | [Changelog](./CHANGELOG.md) | Release history and current release status |
  305. | [Installation Guide](./docs/installation.md) | Detailed setup instructions |
  306. | [Runtime Targets](./docs/runtime-targets.md) | Claude Code and Codex installation targets |
  307. | [Showcase](./docs/showcase.md) | Copy-paste workflow demos |
  308. | [Spellbook Operating Contract](./docs/spellbook-operating-contract.md) | Agent behavior rules for autonomy, escalation, pushback, feedback loops, and done-when checks |
  309. | [Skill Format Policy](./docs/skill-format-policy.md) | Directory vs file skill layout rules |
  310. | [Skill Quality Playbook](./docs/skill-quality-playbook.md) | Trigger descriptions, gotchas, progressive disclosure, and verification |
  311. | [Skill Testing Guide](./docs/skill-testing-guide.md) | How to validate skills work |
  312. | [Creating Plugins](./docs/creating-plugins.md) | Build your own skills |
  313. | [Product Lifecycle (EN)](./docs/product-lifecycle-skills-en.md) | Full lifecycle coverage |
  314. | [Product Lifecycle (中文)](./docs/product-lifecycle-skills-zh.md) | 产品生命周期覆盖 |
  315. ---
  316. ## Release Status
  317. Spellbook is in pre-1.0 release-readiness mode. The first numbered tag is
  318. [v0.1.0](https://github.com/majiayu000/spellbook/releases/tag/v0.1.0). The
  319. install path remains the repository `main` branch.
  320. See [Changelog](./CHANGELOG.md) for release history.
  321. Current limitations:
  322. - Codex installs skills only; Claude Code agents are skipped for Codex targets.
  323. - Some skills depend on external CLIs, accounts, credentials, or platform access
  324. that are not bundled by the installer.
  325. - The registry validator checks installable skill structure, not every external
  326. workflow end to end.
  327. Support paths:
  328. - Bugs: [open an issue](https://github.com/majiayu000/spellbook/issues/new/choose)
  329. - Feature ideas: [open a feature request](https://github.com/majiayu000/spellbook/issues/new/choose)
  330. - Security vulnerabilities: follow [Security Policy](./SECURITY.md)
  331. ---
  332. ## Credits
  333. Built on the shoulders of giants:
  334. - [anthropics/skills](https://github.com/anthropics/skills) - Official Anthropic skills
  335. - [obra/superpowers](https://github.com/obra/superpowers) - Development methodology
  336. - [claude-code-plugins-plus](https://github.com/jeremylongshore/claude-code-plugins-plus) - Plugin hub
  337. - [VoltAgent/awesome-claude-code-subagents](https://github.com/VoltAgent/awesome-claude-code-subagents) - Agent collection
  338. ---
  339. ## Contributing
  340. Contributions welcome! Please read our [Contributing Guide](./CONTRIBUTING.md) first.
  341. - Found a bug? [Open an issue](https://github.com/majiayu000/spellbook/issues)
  342. - Have a skill idea? [Open a feature request](https://github.com/majiayu000/spellbook/issues/new/choose)
  343. - Want to contribute? [Submit a PR](https://github.com/majiayu000/spellbook/pulls)
  344. ---
  345. ## The Agent Infra Stack
  346. This project is one layer of an open-source stack for running coding agents (Claude Code, Codex) as serious infrastructure. Every piece works standalone; together they close the loop:
  347. `spellbook` sits in the **Extend** layer — the authoring side of the skill story: write once, run on Claude Code and Codex. Discovery and distribution live in `claude-skill-registry`.
  348. | Layer | Project | What it does |
  349. |---|---|---|
  350. | Extend | [claude-skill-registry](https://github.com/majiayu000/claude-skill-registry) | Discover and search community Claude Code skills |
  351. | Extend | [spellbook](https://github.com/majiayu000/spellbook) **◀ you are here** | Cross-runtime skills for Claude Code, Codex, and multi-agent workflows |
  352. | Trust | [argus](https://github.com/majiayu000/argus) | Static install-time scanner for supply-chain attacks (npm / PyPI / crates.io) |
  353. | Trust | [vibeguard](https://github.com/majiayu000/vibeguard) | Rules, hooks, and guards against hallucinated or unverified agent changes |
  354. | Remember | [remem](https://github.com/majiayu000/remem) | Local-first persistent memory for Claude Code and Codex sessions |
  355. | Orchestrate | [harness](https://github.com/majiayu000/harness) | Rust agent orchestration platform — rules, skills, GC, observability |
  356. | Route | [litellm-rs](https://github.com/majiayu000/litellm-rs) | High-performance Rust AI gateway — 100+ LLM APIs via OpenAI format |
  357. | Keep | [fibo](https://github.com/majiayu000/fibo) | Session command center — monitor, recover, never lose agent work |
  358. ---
  359. ## License
  360. [MIT License](./LICENSE) - Use freely in your projects.
  361. ---
  362. <div align="center">
  363. <p>If this helps you, consider giving it a ⭐</p>
  364. <p>Made for builders using Claude Code, Codex, and multi-agent workflows</p>
  365. </div>
skills/cliproxy-newapi-stack/scripts/verify_stack.sh · Bundled script · 141 lines · 1 finding marked
  1. #!/usr/bin/env bash
  2. # Verify CLIProxyAPI through loopback, NewAPI through HTTPS, and exact billing.
  3. #
  4. # Usage:
  5. # SSH_TARGET=root@1.2.3.4 SSH_KEY=~/.ssh/id_ed25519 \
  6. # CLIPROXY_URL=http://127.0.0.1:8317 NEWAPI_URL=https://api.example.com \
  7. # CLIPROXY_KEY=cpa_xxx NEWAPI_TOKEN=sk-xxx MODEL=gpt-5.4 \
  8. # INPUT_USD_PER_M=2.5 OUTPUT_USD_PER_M=15 QUOTA_PER_UNIT=500000 \
  9. # ./verify_stack.sh
  10. set -euo pipefail
  11. SCRIPT_DIR="$(cd -- "$(dirname -- "${BASH_SOURCE[0]}")" && pwd)"
  12. # shellcheck source=skills/cliproxy-newapi-stack/scripts/lib/validation.sh
  13. . "$SCRIPT_DIR/lib/validation.sh"
  14. : "${SSH_TARGET:?SSH_TARGET required}"
  15. : "${SSH_KEY:=$HOME/.ssh/id_ed25519}"
    • Note: Reads a credential file in a script that also sends data over the network
    • Line 18 on GitHub
  16. : "${CLIPROXY_URL:=http://127.0.0.1:8317}"
  17. : "${NEWAPI_URL:?NEWAPI_URL required and must use HTTPS}"
  18. : "${CLIPROXY_KEY:?CLIPROXY_KEY required}"
  19. : "${NEWAPI_TOKEN:?NEWAPI_TOKEN required}"
  20. : "${MODEL:?MODEL required}"
  21. : "${INPUT_USD_PER_M:?INPUT_USD_PER_M required}"
  22. : "${OUTPUT_USD_PER_M:?OUTPUT_USD_PER_M required}"
  23. : "${QUOTA_PER_UNIT:=500000}"
  24. : "${MAX_QUOTA_ERROR_RATIO:=0.05}"
  25. : "${DB:=/root/newapi/data/one-api.db}"
  26. require_safe_ssh_target SSH_TARGET "$SSH_TARGET"
  27. require_loopback_http_origin CLIPROXY_URL "$CLIPROXY_URL"
  28. require_https_origin NEWAPI_URL "$NEWAPI_URL"
  29. require_safe_token CLIPROXY_KEY "$CLIPROXY_KEY"
  30. require_safe_token NEWAPI_TOKEN "$NEWAPI_TOKEN"
  31. require_safe_model_name MODEL "$MODEL"
  32. require_positive_decimal INPUT_USD_PER_M "$INPUT_USD_PER_M"
  33. require_positive_decimal OUTPUT_USD_PER_M "$OUTPUT_USD_PER_M"
  34. require_positive_uint QUOTA_PER_UNIT "$QUOTA_PER_UNIT"
  35. require_positive_decimal MAX_QUOTA_ERROR_RATIO "$MAX_QUOTA_ERROR_RATIO"
  36. require_absolute_safe_path DB "$DB"
  37. tmp_dir="$(mktemp -d)"
  38. chmod 700 "$tmp_dir"
  39. trap 'rm -rf "$tmp_dir"' EXIT
  40. nonce="$(date +%s)-$$"
  41. payload='{"model":"'"$MODEL"'","messages":[{"role":"user","content":"billing probe '"$nonce"'; reply OK"}],"stream":false,"max_tokens":4}'
  42. probe() {
  43. local base_url=$1 token=$2 output=$3 config=$4
  44. umask 077
  45. {
  46. printf 'silent\nshow-error\n'
  47. printf 'output = "%s"\n' "$output"
  48. printf 'write-out = "%%{http_code}"\n'
  49. printf 'request = "POST"\n'
  50. printf 'url = "%s/v1/chat/completions"\n' "$base_url"
  51. printf 'header = "Authorization: Bearer %s"\n' "$token"
  52. printf 'header = "Content-Type: application/json"\n'
  53. } > "$config"
  54. curl --config "$config" --data-binary @- <<< "$payload"
  55. rm -f "$config"
  56. }
  57. echo "[1/3] Direct CLIProxyAPI through loopback tunnel"
  58. http_a="$(probe "$CLIPROXY_URL" "$CLIPROXY_KEY" "$tmp_dir/cpa.json" "$tmp_dir/cpa.curl")"
  59. echo " HTTP $http_a"
  60. [ "$http_a" = "200" ] || {
  61. echo "FAIL direct CLIProxyAPI"
  62. cat "$tmp_dir/cpa.json"
  63. exit 1
  64. }
  65. echo "[2/3] Via NewAPI HTTPS"
  66. http_b="$(probe "$NEWAPI_URL" "$NEWAPI_TOKEN" "$tmp_dir/newapi.json" "$tmp_dir/newapi.curl")"
  67. echo " HTTP $http_b"
  68. [ "$http_b" = "200" ] || {
  69. echo "FAIL via NewAPI"
  70. cat "$tmp_dir/newapi.json"
  71. exit 1
  72. }
  73. request_id="$(python3 -c 'import json,sys; print(json.load(open(sys.argv[1], encoding="utf-8")).get("id", ""))' "$tmp_dir/newapi.json")"
  74. require_safe_token request_id "$request_id"
  75. echo "[3/3] Verify matching NewAPI billing log"
  76. ssh -i "$SSH_KEY" -- "$SSH_TARGET" python3 - \
  77. "$DB" "$request_id" "$MODEL" "$INPUT_USD_PER_M" "$OUTPUT_USD_PER_M" \
  78. "$QUOTA_PER_UNIT" "$MAX_QUOTA_ERROR_RATIO" <<'PY'
  79. from decimal import Decimal
  80. import sqlite3
  81. import sys
  82. (
  83. db_path,
  84. request_id,
  85. expected_model,
  86. input_usd_per_m,
  87. output_usd_per_m,
  88. quota_per_unit,
  89. max_error_ratio,
  90. ) = sys.argv[1:]
  91. connection = sqlite3.connect(f"file:{db_path}?mode=ro", uri=True)
  92. try:
  93. row = connection.execute(
  94. "SELECT model_name, prompt_tokens, completion_tokens, quota "
  95. "FROM logs WHERE request_id = ? ORDER BY id DESC LIMIT 1",
  96. (request_id,),
  97. ).fetchone()
  98. finally:
  99. connection.close()
  100. if row is None:
  101. raise RuntimeError("matching billing log not found")
  102. model_name, prompt_tokens, completion_tokens, quota = row
  103. if model_name != expected_model:
  104. raise RuntimeError(
  105. f"billing log model mismatch: expected {expected_model!r}, got {model_name!r}"
  106. )
  107. if prompt_tokens is None or completion_tokens is None or quota is None:
  108. raise RuntimeError(f"billing log has null accounting fields: {row!r}")
  109. error_ratio = Decimal(max_error_ratio)
  110. if error_ratio > Decimal(1):
  111. raise RuntimeError("MAX_QUOTA_ERROR_RATIO must not exceed 1")
  112. expected_quota = (
  113. Decimal(prompt_tokens) * Decimal(input_usd_per_m)
  114. + Decimal(completion_tokens) * Decimal(output_usd_per_m)
  115. ) * Decimal(quota_per_unit) / Decimal(1_000_000)
  116. actual_quota = Decimal(quota)
  117. allowed_error = max(Decimal(1), expected_quota * error_ratio)
  118. if abs(actual_quota - expected_quota) > allowed_error:
  119. raise RuntimeError(
  120. "billing quota mismatch: "
  121. f"expected {expected_quota}, got {actual_quota}, tolerance {allowed_error}"
  122. )
  123. print("model_name\tprompt_tokens\tcompletion_tokens\tquota")
  124. print("\t".join(map(str, row)))
  125. PY
  126. echo "OK both paths returned 200 and the matching billing quota is within tolerance"

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.