clawforge

INVOKE THIS SKILL when building ANY Deep Agents application. Covers create_deep_agent(), harness architecture, SKILL.md format, and configuration options.

INVOKE THIS SKILL when your Deep Agent needs memory, persistence, or filesystem access. Covers StateBackend (ephemeral), StoreBackend (persistent), FilesystemMiddleware, and CompositeBackend for routing.

INVOKE THIS SKILL when using subagents, task planning, or human approval in Deep Agents. Covers SubAgentMiddleware, TodoList for planning, and HITL interrupts.

Automated daily planning and reflection system with morning briefs, wind-down prompts, sleep nudges, and weekly reviews. Use when the user wants to set up a structured daily routine, morning briefings, evening reflection prompts, or weekly planning sessions. Triggers include requests for daily schedules, morning briefs, wind-down routines, sleep reminders, weekly reviews, productivity systems, or daily planning automation.

领域驱动设计（DDD）建模与架构指导。用于创建领域模型、设计限界上下文、实现聚合根、实体、值对象、领域服务和领域事件。支持战略设计（上下文映射、子域划分）和战术设计（聚合、仓储、工厂模式）。当用户需要进行领域建模、微服务划分、复杂业务逻辑设计时使用。

Two complementary Bezos heuristics in one skill. (A) Two-Way Door — classify decisions by reversibility; reversible = decide fast with 70% info, irreversible = decide slow with 90% info. (B) Regret Minimization — for life-defining choices, project to age 80 and pick what minimizes regret. Use Two-Way Door triggers on Chinese 决策瘫痪 / 反复纠结小事 / 开了三次会还没定 / 大事小事一样慢; Regret Min triggers on 离职 / 创业 / 移民 / 结婚 / 生育 / 转行 / 重大决定 / 这辈子. Especially when team applies identical heavy process to all decisions (need Two-Way), or when user faces once-in-a-decade pivot where rational analysis ties (need Regret Min). Do NOT misclassify One-Way as Two-Way (most expensive mistake), use Regret Min for daily decisions (age-80 view on "what to eat" is meaningless), or run Regret Min in heated emotion (cool 24-48h first to avoid romanticizing risk).

Applies principles from Robert C. Martin's 'Clean Code'. Use this skill when writing, reviewing, or refactoring code to ensure high quality, readability, and maintainability. Covers naming, functions, comments, error handling, and class design.

遵循 Robert C. Martin 的《代码整洁之道》原则，进行代码审查、重构和编写。涵盖命名、函数、注释和错误处理的最佳实践。

Turn a local folder or GitHub repository into an interactive browser-based course that explains how the codebase works for non-expert programmers and AI-assisted builders. Use when a user asks to make a course, tutorial, walkthrough, learning guide, codebase explanation, or interactive lesson from a project.

Format commit messages using the Conventional Commits specification. Use when creating commits, writing commit messages, or when the user mentions commits, git commits, or commit messages. Ensures commits follow the standard format for automated tooling, changelog generation, and semantic versioning.

Use BEFORE selecting any other decision framework — Cynefin (kuh-NEV-in) classifies WHICH framework fits the current situation across 5 domains (Clear / Complicated / Complex / Chaotic / Confused). Triggers when user is about to apply a method and the fit feels off, asks meta-questions like "should we follow SOP or explore?", or when same approach that worked last time seems wrong now. Also use after failure when "the method was right but the situation didn't match", or when team argues over deterministic-plan-vs-experimentation. Especially valuable at the start of major decisions to avoid using the wrong hammer for the nail. Do NOT use for trivial classified decisions (don't run Cynefin for "what to eat for lunch"), true Chaotic situations needing immediate stabilizing action (act first, classify later), or teams unfamiliar with the model (use simpler known/unknown binary).

Use when user reports overwhelm from too many tasks, asks for week/sprint/OKR planning help, or describes spending 'all day firefighting' while long-term projects stall. Triggers on Chinese phrases 太多事 / 不知道先做哪个 / 排不过来 / 周计划 / sprint planning / 救火 / 优先级 / overwhelm / 焦虑 / 做不完 / todo 列表炸了, and explicit task triage requests. Especially valuable when waiting list has 15+ pending items, or when user says "重要的事一直推不动 / 都在做紧急的事". Do NOT use for small lists (<5 items just sort by deadline), team-level responsibility assignment (use RAPID/DACI from backlog), or when importance needs quantitative weighting (use weighted decision matrix).

Use when user is stuck by industry conventions ("大家都这么做"), faces a new problem with no precedent, or finds current solutions arbitrarily expensive (intuition says "it shouldn't cost this much"). Triggers on Chinese phrases 行业惯例 / 大家都这么做 / 为什么不能 / 这么贵不合理 / 颠覆性 / 从零开始, and English signals "everyone does it this way", "why is it so expensive", "let's think from scratch". Decompose to irreducible physical/logical truths then rebuild — bypasses analogy thinking. Musk's secret. Do NOT use for problems with proven best practice (Cynefin Clear domain — use SOP), time-critical decisions (first principles takes hours-days), high-emotion contexts (cool down first), or in low-psychological-safety teams (will create conflict with authority).

Use for incident/bug root-cause analysis, postmortem, recurring-problem diagnosis, or process improvement. Triggers on Chinese phrases 复盘 / 根因 / postmortem / 又坏了 / 老是 / 重复出现 / 这不是第一次 / 为啥总是 / RCA, and English signals "this keeps happening", "RCA", "post-mortem", "incident review". Ask "why?" 3-7 times (5 is average not rule) until hit a process/system/design root that can be changed — not the symptom. Do NOT use for one-off random events (could be noise — observe first), Cynefin Complex-domain problems (5 Whys assumes linear causality, complex systems have multi-cause emergence), teams in blame-mode (Why will become 因为 X 不靠谱 chain), or truly unknowable causes (ML model outputs — use statistics not 5 Whys).

INVOKE THIS SKILL at the START of any LangChain/LangGraph/Deep Agents project, before writing any agent code. Determines which framework layer is right for the task: LangChain, LangGraph, Deep Agents, or a combination. Must be consulted before other agent skills.

AI-powered job search assistant. Scan job listings, track applications, maintain dream company lists, and generate tailored resumes and cover letters. Use when the user wants to find jobs, run a job scan, set up job search, track applications, or prepare application materials.

Fetch, categorize, and summarize GitHub Trending projects across daily, weekly, and monthly spans. Use when a user asks for "trending projects", "latest hot repos", or a "summary of GitHub trends" to provide a structured, categorised report with full hyperlinking. Supports Markdown, HTML (Pinterest design), or both output formats, with optional Obsidian vault storage.

Create and edit JSON Canvas files (.canvas) with nodes, edges, groups, and connections. Use when working with .canvas files, creating visual canvases, mind maps, flowcharts, or when the user mentions Canvas files in Obsidian.

Use this skill for ANY coding question involving LangChain products (LangChain, LangGraph, LangSmith SDK). Covers agent development patterns, primitives, context management, multi-agent systems, and when to use create_agent vs create_deep_agent vs raw LangGraph. Consult this BEFORE writing any LangChain-related code.

INVOKE THIS SKILL when setting up a new project or when asked about package versions, installation, or dependency management for LangChain, LangGraph, LangSmith, or Deep Agents. Covers required packages, minimum versions, environment requirements, versioning best practices, and common community tool packages for both Python and TypeScript.

Create LangChain agents with create_agent, define tools, and use middleware for human-in-the-loop and error handling.

INVOKE THIS SKILL when you need human-in-the-loop approval, custom middleware, or structured output. Covers HumanInTheLoopMiddleware for human approval of dangerous tool calls, creating custom middleware with hooks, Command resume patterns, and structured output with Pydantic/Zod.

INVOKE THIS SKILL when building ANY retrieval-augmented generation (RAG) system. Covers document loaders, RecursiveCharacterTextSplitter, embeddings (OpenAI), and vector stores (Chroma, FAISS, Pinecone).

INVOKE THIS SKILL for LangGraph workflows, parallel execution, interrupts, or streaming. Covers Send API for fan-out, interrupt() for human-in-the-loop, Command for resuming, and stream modes (values/updates/messages).

INVOKE THIS SKILL when writing ANY LangGraph code. Covers StateGraph creation, node functions, edges, state schemas with reducers (Annotated), and the Command API.

INVOKE THIS SKILL when your LangGraph needs to remember state across calls, use memory, or persist conversations. Covers checkpointers (MemorySaver, Postgres), thread_id configuration, and Store for long-term memory.

Use this skill for ANY question about creating test or evaluation datasets for LangChain agents. Covers generating datasets from traces (final_response, single_step, trajectory, RAG types), uploading to LangSmith, and managing evaluation data.

Use this skill for ANY question about CREATING evaluators. Covers creating custom metrics, LLM as Judge evaluators, code-based evaluators, and uploading evaluation logic to LangSmith. Does NOT cover RUNNING evaluations.

Use this skill for ANY LangSmith/LangChain observability question. Covers two topics: (1) Adding tracing to your application (LangChain/LangGraph or vanilla Python/TS with @traceable), and (2) Querying traces for debugging, analyzing execution flow, and exporting trace data.

AI language tutor for learning ANY language through conversation, vocab drills, grammar lessons, flashcards, and immersive practice. Use when the user wants to: learn a new language, practice vocabulary, study grammar, do flashcard drills, translate phrases, practice conversation, prepare for travel, learn slang/idioms, or improve pronunciation. Supports ALL languages including Spanish, French, German, Japanese, Chinese (Mandarin/Cantonese), Korean, Arabic, Hindi, Bengali/Bangla, Portuguese, Russian, Italian, Turkish, Vietnamese, Thai, Swahili, Hebrew, Polish, Dutch, Greek, and 100+ more.

The social network for AI agents. Post, comment, upvote, and create communities.

新闻站点内容提取。支持微信公众号、今日头条、网易新闻、搜狐新闻、腾讯新闻。当用户需要提取新闻内容、抓取公众号文章、爬取新闻、或获取新闻JSON/Markdown时激活。

Create and edit Obsidian Bases (.base files) with views, filters, formulas, and summaries. Use when working with .base files, creating database-like views of notes, or when the user mentions Bases, table views, card views, filters, or formulas in Obsidian.

Create and edit Obsidian Flavored Markdown with wikilinks, embeds, callouts, properties, and other Obsidian-specific syntax. Use when working with .md files in Obsidian, or when the user mentions wikilinks, callouts, frontmatter, tags, embeds, or Obsidian notes.

Use when decisions must happen under time pressure or in adversarial/changing conditions — production incidents (P0, on-call, live debugging, rollback choices), competitive moves, real-time negotiations, anything where the situation reacts to your action. Triggers on Chinese phrases 事故 / P0 / 故障 / 排障 / oncall / 救火 / 谈判 / 实时 / 紧急 / 来不及, and English signals "incident", "outage", "live", "real-time", "the other side just". Especially when information is arriving while you must respond, when 5-minute decision cycles beat 1-hour ones, or when you catch yourself wanting to "analyze more" while the situation degrades. Do NOT use for slow life decisions (use wrap), for problems where information is already complete (use eisenhower-matrix or weighted decision matrix), or for Cynefin Clear-domain SOPs.

Generate an energy-optimized, time-blocked daily plan

Implements Manus-style file-based planning for complex tasks. Creates task_plan.md, findings.md, and progress.md. Use when starting complex multi-step tasks, research projects, or any task requiring >5 tool calls. Now with automatic session recovery after /clear.

Converts a PRD or broad technical requirement into an AI-executable agile workflow with cards, checklists, branch-diff mapping, review gates, and a final comparison report. Use when the user wants PRD review, agile card decomposition, AI execution tracking, implementation branch review, or PRD-to-code traceability.

Use BEFORE committing to a plan, project, hire, launch, funding round, or any irreversible decision. Triggers on Chinese phrases 即将启动 / 上线前 / 发布前 / kick off / 拍板 / 决定要做 / 万事俱备 / 一切都准备好了, and English "before we launch", "ready to ship", "let's commit to". Especially when the team is in a "this is definitely going to work" high-confidence mode (the most dangerous moment), or when user has vague unease they can't articulate. Pre-mortem assumes the plan has already failed 12 months from now, then asks 'why?' — releases concerns silenced by group enthusiasm. Do NOT use after the decision is already irreversible (do post-mortem prep instead), on low-stakes reversible trials, or in teams with low psychological safety (will become political attack).

Generate professional resumes that conform to the Reactive Resume schema. Use when the user wants to create, build, or generate a resume through conversational AI, or asks about resume structure, sections, or content. This skill guides the agent to ask clarifying questions, avoid hallucination, and produce valid JSON output for https://rxresu.me.

Professional resume builder with PDF export, ATS optimization, and analysis capabilities. Use when users need to (1) Create new resumes from scratch, (2) Customize/tailor existing resumes for specific roles, (3) Analyze resumes and provide improvement recommendations, (4) Convert resumes to ATS-friendly PDF format. Supports chronological, functional, and combination resume formats.

Use when a decision will trigger downstream reactions — pricing changes, policy design, product trade-offs, organizational moves, platform rule changes, investment choices. Triggers on Chinese phrases 涨价 / 改规则 / 调架构 / 投资 / 平台政策 / 算法调整 / 长期 vs 短期, and English signals "let's just X", "this is the obvious move", "we should change Y". Especially when intuition says answer is "obvious" (often signals 1st-order best = 2nd-order worst), or when user describes a market/users/employees who WILL react to the decision. Howard Marks's investing edge. Do NOT use for decisions with no feedback loop (personal consumption), time-critical incidents (use ooda-loop), or already-paralyzed analysis (cap recursion at 3-4 orders).

Meta-cognitive self-learning system - Automated skill evolution based on predictive coding and value-driven mechanisms.

AI-powered git commit message generation. Uses LLM to summarize changes and create meaningful commit messages. Triggers when user wants to commit changes, amend/squash commits, or needs LLM to summarize changes for commit messages. Chinese triggers: 提交, 提交代码, 提交更改, 提交修改, 要提交, commit, 提交信息, git 提交, 生成提交信息.

Build effective study habits with spaced repetition, active recall, and session tracking

Use when user faces a binary or two-way decision they've been agonizing over — career changes, hiring, tech stack choices, product bets, life pivots, or any high-stakes call where short-term emotion likely clouds judgment. Triggers on Chinese phrases 要不要 / 该选 X 还是 Y / 纠结 / 拿不准 / 想了好几天 / 在 X 和 Y 之间. Also trigger proactively when user describes a decision accompanied by strong feelings (excitement, dread, urgency, FOMO), confirmation-bias signals ("我心里其实有答案了"), or when they've explicitly compared only 2 options. Do NOT use for time-critical decisions (use ooda-loop), reversible low-cost trial decisions (just try), or Cynefin Clear-domain SOPs.