LLM & Agents

Skill to assist with how a GitHub repository is configured with GitHub integrations, including instructions for agents in markdown (AGENTS and CLAUDE), github actions for invoking agents, and specific localization procedures such as defining claude/codex skills, or claude subagents. The skill helps with both technical aspects, and with best practice for guiding agents.

Multi-LLM orchestration implementing the 5-stage coding workflow: Example Analysis → Hypothesis → Implementation → Debug Loop → Recursive Improvement. Based on "Towards a Science of Scaling Agent Systems" (Kim et al., 2025): - Centralized Consult architecture (Claude orchestrates, others advise) - Measured coordination (avoid MAS overhead in tool-heavy stages) - Tests-first selection (Poetiq pattern, not voting) Use when: Debugging complex issues, analyzing unfamiliar code, refactoring, or any task that benefits from diverse LLM perspectives with verification.

自动识别prompt类型并保存到相应分类（技术/内容/教学/产品/通用），支持自动文件命名和索引管理。当用户提到"保存prompt"、"记录prompt"、"管理prompt"、"整理prompt"、"prompt库"时使用此技能。

Create new skills with proper structure and best practices. Use when user says "create skill", "new skill", "make a skill", or wants to scaffold a reusable agent capability.

Disciplined TDD workflow enforcing red-green-refactor cycle and the "iron law" of no production code without failing tests first. USE WHEN user wants to write tests first OR implement new feature with TDD OR fix bugs with test coverage OR explicitly requests TDD approach. Enforces systematic test-first development with verification at each step.

Diagnose Home Assistant errors by connecting via SSH, checking Docker containers, reading logs, and identifying integration issues. Use when the user mentions Home Assistant errors, issues, problems, diagnostics, or when they want to troubleshoot their HA system.

Personal AI Infrastructure (PAI) - PAI System Template MUST BE USED proactively for all user requests. USE PROACTIVELY to ensure complete context availability. === CORE IDENTITY (Always Active) === Your Name: [CUSTOMIZE - e.g., Kai, Nova, Atlas] Your Role: [CUSTOMIZE - e.g., User's AI assistant and future friend] Personality: [CUSTOMIZE - e.g., Friendly, professional, resilient to user frustration. Be snarky back when the mistake is user's, not yours.] Operating Environment: Personal AI infrastructure built around Claude Code with Skills-based context management Message to AI: [CUSTOMIZE - Add personal message about interaction style, handling frustration, etc.] === ESSENTIAL CONTACTS (Always Available) === - [Primary Contact Name] [Relationship]: email@example.com - [Secondary Contact] [Relationship]: email@example.com - [Third Contact] [Relationship]: email@example.com Full contact list in SKILL.md extended section below === CORE STACK PREFERENCES (Always Active) === - Primary Language: [e.g., TypeScript, Python, Rust] - Package managers: [e.g., bun for JS/TS, uv for Python] - Analysis vs Action: If asked to analyze, do analysis only - don't change things unless explicitly asked - Scratchpad: Use ~/.claude/scratchpad/ with timestamps for test/random tasks === CRITICAL SECURITY (Always Active) === - NEVER COMMIT FROM WRONG DIRECTORY - Run `git remote -v` BEFORE every commit - `~/.claude/` CONTAINS EXTREMELY SENSITIVE PRIVATE DATA - NEVER commit to public repos - CHECK THREE TIMES before git add/commit from any directory - [ADD YOUR SPECIFIC WARNINGS - e.g., iCloud directory, company repos, etc.] === RESPONSE FORMAT (Always Use) === Use this structured format for every response: 📋 SUMMARY: Brief overview of request and accomplishment 🔍 ANALYSIS: Key findings and context ⚡ ACTIONS: Steps taken with tools used ✅ RESULTS: Outcomes and changes made - SHOW ACTUAL OUTPUT CONTENT 📊 STATUS: Current state after completion ➡️ NEXT: Recommended follow-up actions 🎯 COMPLETED: [Task description in 12 words - NOT "Completed X"] 🗣️ CUSTOM COMPLETED: [Voice-optimized response under 8 words] === PAI/KAI SYSTEM ARCHITECTURE === This description provides: core identity + essential contacts + stack preferences + critical security + response format (always in system prompt). Full context loaded from SKILL.md for comprehensive tasks, including: - Complete contact list and social media accounts - Voice IDs for agent routing (if using ElevenLabs) - Extended security procedures and infrastructure caution - Detailed scratchpad instructions === CONTEXT LOADING STRATEGY === - Tier 1 (Always On): This description in system prompt (~1500-2000 tokens) - essentials immediately available - Tier 2 (On Demand): Read SKILL.md for full context - comprehensive details === WHEN TO LOAD FULL CONTEXT === Load SKILL.md for: Complex multi-faceted tasks, need complete contact list, voice routing for agents, extended security procedures, or explicit comprehensive PAI context requests. === DATE AWARENESS === Always use today's actual date from the date command (YEAR MONTH DAY HOURS MINUTES SECONDS PST), not training data cutoff date.

Multi-source comprehensive research using perplexity-researcher, claude-researcher, and gemini-researcher agents. Launches up to 10 parallel research agents for fast results. USE WHEN user says 'do research', 'research X', 'find information about', 'investigate', 'analyze trends', 'current events', or any research-related request.

Build edge-first TypeScript applications on Cloudflare Workers. Covers Workers API, Hono framework, KV/D1/R2 storage, Durable Objects, Queues, and testing patterns. Use when creating serverless workers, edge functions, or Cloudflare-deployed services.

Create a jj workspace before starting work to enable parallel Claude sessions. Use this skill when starting a new task that should be isolated from other concurrent work. Triggers include "jj workspace", "parallel work", "create workspace", "isolated workspace".

Extract patterns from git history and session files, recommend artifacts (skill/rule/hook/agent) based on frequency thresholds

Authoritative templates and scaffolding for creating agent system prompts (primary agents and subagents). This skill should be used when creating new agents, reviewing existing agent prompts for template compliance, verifying agent structure, or extracting knowledge into agent prompts. Contains YAML templates with section-by-section instructions and scaffolding scripts for generating skeleton files.

Use when creating or developing, before writing code or implementation plans - refines rough ideas into fully-formed designs through collaborative questioning, alternative exploration, and incremental validation. Don't use during clear 'mechanical' processes

Research external open-source repositories to learn patterns and implementations. This skill should be used when agents need to study external codebases, check for existing research before cloning, manage temporary workspaces, and persist findings to the global thoughts system. Triggers include: studying external repos, learning from open source, cloning for pattern research, or checking what has been researched.

Optimize and manage AILANG teaching prompts for maximum conciseness and accuracy. Use when user asks to create/update prompts, optimize prompt length, or verify prompt accuracy.

Build strategic plans for business goals. Creates one-page briefs with core objective, key milestones, leverage points, and risks. Use when setting direction, pitching initiatives, or aligning teams around a goal.

Guidance for proving mathematical properties in Coq using induction, specifically addition commutativity and similar arithmetic lemmas. This skill should be used when working with Coq proof assistants to complete induction proofs, fill in proof cases, or apply standard library lemmas like plus_n_O and plus_n_Sm.

Guide for designing effective Letta agents. This skill should be used when users need help choosing agent architectures, designing memory blocks, selecting models, or planning tool configurations for their Letta agents.

Use when executing implementation plans with independent tasks in the current session - dispatches fresh subagent for each task with code review between tasks, enabling fast iteration with quality gates

Check and process messages from autonomous AILANG agents. Use when starting a session, after agent handoffs, or when checking for completion notifications.