提示词
提示词工程与模板
agents-md-improver
ericmjl
Keeps repo-local agent instructions consistent by proposing updates to AGENTS.md when a user corrects the coding agent or asks to change AGENTS.md, CLAUDE.md, .claude/CLAUDE.md, or GEMINI.md.
publish-to-google-docs
ericmjl
Quickly publish markdown notes to styled Google Docs using pandoc and a Word template. Use when creating Google Docs from markdown, applying branding to documents, or pushing content to Drive for sharing.
devcontainer-setup
shipshitdev
Set up a .devcontainer for VS Code with Docker, Claude Code CLI support, and configurable options. Use when the user asks to "set up devcontainer", "add docker development environment", "configure dev container", or needs to containerize their development workflow.
finance-data
macaron-software
Financial data modeling, calculations, and transaction categorization for personal wealth management. Use when working on database schema, P&L calculations, asset allocation, dividend tracking, net worth computation, or automatic transaction categorization with regex rules and LLM fallback.
researching-web
alexei-led
Web research via Perplexity AI. Use for technical comparisons (X vs Y), best practices, industry standards, documentation. Triggers on "research", "compare", "vs", "best practice", "which is better", "pros and cons".
advanced-evaluation
shipshitdev
Master LLM-as-a-Judge evaluation techniques including direct scoring, pairwise comparison, rubric generation, and bias mitigation. Use when building evaluation systems, comparing model outputs, or establishing quality standards for AI-generated content.
observe
saadshahd
On-demand codebase health check. Use when "observe", "health check", "is the codebase healthy", "check for drift", "system status", or before starting a large feature.
recipe-generator
cclank
自动生成高品质菜谱的技能。当用户需要创建详细的菜谱、烹饪教程、食谱配方时使用。输入菜品名称后，检索权威来源（米其林、舌尖上的中国等）的制作工艺，生成包含完整准备工作、详细制作步骤、注意事项和专业技巧的菜谱文档，并最终生成传统中式风格的菜谱海报生图提示词。
maintaining-project-context
ed3dai
Use when completing development phases or branches to identify and update CLAUDE.md or AGENTS.md files that may have become stale - analyzes what changed, determines affected contracts and documentation, and coordinates updates
writing-skills
ed3dai
Use when creating new skills, editing existing skills, or verifying skills work before deployment - applies TDD to process documentation by testing with subagents before writing, iterating until bulletproof against rationalization
writing-claude-md-files
ed3dai
Use when creating or updating CLAUDE.md files for projects or subdirectories - covers top-level vs domain-level organization, capturing architectural intent and contracts, and mandatory freshness dates
using-git-worktrees
ed3dai
Use when starting feature work that needs isolation from current workspace or before executing implementation plans - creates isolated git worktrees with smart directory selection and safety verification
bootstrap
geekatron
This skill should be used when the user says "bootstrap", "set up Jerry", "configure guardrails", "initialize Jerry", or asks how to get started with Jerry. Sets up Jerry's context distribution by syncing behavioral rules and patterns.
add-skill
htlin222
Create new Claude Code skills with proper structure. Use when user wants to add a skill, create a slash command, or automate a workflow. Fetches official docs and guides through skill creation.
adversary
geekatron
On-demand adversarial quality reviews using strategy templates. Selects strategies by criticality level, executes adversarial templates against deliverables, and scores quality using LLM-as-Judge rubric. Integrates with quality-enforcement.md SSOT.
ux-behavior-design
geekatron
"Fogg Behavior Model B=MAP bottleneck diagnosis sub-skill for the /user-experience parent skill. Diagnoses why users fail to take desired actions by analyzing the three B=MAP factors (Motivation, Ability, Prompt) and identifying which factor falls below the action threshold. Produces bottleneck diagnoses, factor-level assessments, and intervention recommendations with synthesis confidence gates. Invoke when teams need to understand why users are not completing a specific action, diagnose behavioral bottlenecks, design behavior change interventions, or analyze post-launch user inaction patterns. Invoked by ux-orchestrator during Wave 4 lifecycle-stage routing or when user intent is \"Users not completing action\" during the \"After launch\" stage. Triggers: behavior design, B=MAP, Fogg model, behavior bottleneck, motivation analysis, ability analysis, prompt design, why users don't, user inaction, behavior diagnosis, tiny habits, action threshold."
cpp
htlin222
Write modern C++ with RAII, smart pointers, and STL. Use for C++ development, memory safety, or performance optimization.
sentient
htlin222
Manage, organize and ships your codebase. Use to test Claude Code hook blocking for dangerous commands.
diataxis
geekatron
"Four-quadrant documentation framework. Produces tutorials (learning by doing), how-to guides (goal-oriented tasks), reference documentation (authoritative description), and explanation (conceptual understanding) using the Diataxis methodology. Invoke when creating new documentation, auditing existing docs for quadrant mixing, or classifying documentation requests. Triggers: documentation, tutorial, how-to, howto, reference docs, explanation, diataxis, write docs, write documentation, write tutorial, create documentation, classify documentation, audit documentation, user guide, getting started, quickstart, API docs, developer guide, quadrant, doc type, how-to guide."
checkpoint
htlin222
Git checkpoint management - create, list, restore checkpoints. Invoke when starting risky changes or needing safe recovery points.
working-with-claude-code
obra
Use when working with Claude Code CLI, plugins, hooks, MCP servers, skills, configuration, or any Claude Code feature - provides comprehensive official documentation for all aspects of Claude Code
web-auto-form
DUZ1287
JSON 驱动的浏览器表单自动化工具，为 AI Agent 提供原生 function-calling 集成，支持表单填写、条件分支、数据提取与 PII 脱敏
self-improvement
jdrhyne
"Captures learnings, errors, and corrections to enable continuous improvement. Use when: (1) A command or operation fails unexpectedly, (2) User corrects Claude ('No, that's wrong...', 'Actually...'), (3) User requests a capability that doesn't exist, (4) An external API or tool fails, (5) Claude realizes its knowledge is outdated or incorrect, (6) A better approach is discovered for a recurring task. Also review learnings before major tasks."
codex
jdrhyne
Use when the user asks to run Codex CLI (codex exec, codex resume) or references OpenAI Codex for code analysis, refactoring, or automated editing. Uses GPT-5.2 by default for state-of-the-art software engineering.