LLM & Agents

Search and analyze Claude Code conversation history. Use when user asks about past conversations, previous solutions, what was discussed earlier, finding something from history, or analyzing usage patterns. Triggers include "what did we discuss", "find that conversation", "search history", "past sessions", "how much time", "token usage", "which tools".

Call LLM via CLI for transcription, vision, speech/image generation, piping prompts, sub-agents, ...

Advanced TypeScript patterns for type-safe, maintainable code using sophisticated type system features. Use when building type-safe APIs, implementing complex domain models, or leveraging TypeScript's advanced type capabilities.

Use when noticing something off (code smell, design smell, workaround) while working and want to capture it without losing flow - grounds observation in codebase and files a proper bead that future agents can execute

Creates and updates Claude Code skills following best practices. Guides through requirements gathering, design, content generation, and validation. Handles both new skill creation and refactoring existing skills. Use when user mentions 'create skill', 'build skill', 'new skill', 'update skill', 'improve skill', 'refactor skill', or asks to build capabilities for Claude Code.

Markdown documentation standards for LLM-optimized architecture including section numbering rules for Pandoc PDF generation. Use when writing markdown documentation, creating skills, or authoring content that may be converted to PDF.

Generate comprehensive, publication-quality technical manuals with thematic storytelling using multi-agent orchestration. Use when user asks for themed documentation, narrative technical guides, memorable training materials, or mentions themes like Victorian, steampunk, art deco, medieval, or nautical combined with technical topics. Triggers on phrases like "create themed manual", "documentation with storytelling", or requests to make docs "more engaging" or "memorable".

Generates Jest or Pytest tests following Ben's testing standards. Use when creating tests, adding test coverage, writing unit tests, mocking dependencies, or when user mentions testing, test cases, Jest, Pytest, fixtures, assertions, or coverage.

When setting up commit message validation for a project. When project has commitlint.config.js or .commitlintrc files. When configuring CI/CD to enforce commit format. When extracting commit rules for LLM prompt generation. When debugging commit message rejection errors.

This skill is for writing integrations to the Ruby SDK. Claude acts as the engineer implementing LLM provider or agentic framework integrations. Use when adding support for OpenAI-like providers, Anthropic-like providers, or agent frameworks. Covers TDD workflow, comprehensive testing (streaming/non-streaming/tokens/multimodal), defensive coding, MCP validation, and StandardRB compliance.

Build Model Context Protocol (MCP) servers - comprehensive coverage of generic MCP protocol AND FastMCP framework specialization. Use when creating any MCP server (Python FastMCP preferred, TypeScript/Node also covered). Includes agent-centric design principles, evaluation creation, Pydantic/Zod validation, async patterns, STDIO/HTTP/SSE transports, FastMCP Cloud deployment, .mcpb packaging, security patterns, and mid-2025+ community practices. Standalone skill with no external dependencies.

Configure Ralph loop guidance via AUQ flows - forbidden/encouraged items from constraint scan. TRIGGERS - session guidance, loop configuration, /ralph:start Step 1.6.

Training manager for Hugging Face Jobs - launch fine-tuning on HF cloud GPUs with optional WandB monitoring

Autonomously execute ALL tasks from the backlog until complete. Runs in a continuous loop - picks task, implements via subagent, moves to done, repeats. No user prompts between tasks. Follows task-board workflow (backlog → in-progress → done).

Creates and updates Agent Skills (SKILL.md directories) with proper YAML frontmatter, progressive disclosure patterns, and tool restrictions. Use when creating skills, authoring Agent Skills, building reusable capabilities, or when users mention skills, SKILL.md, skill development, agent capabilities, or progressive disclosure.

Website sitemap generation workflow. Analyzes requirements, delegates to sitemap-analyst agent, generates sitemap with shadcn-ui-blocks component selections. Use when planning website architecture, structuring pages, selecting components, or requesting sitemap/page structure planning.

Systematically analyze a Magento 1 module to determine its purpose, usage, and migration requirements for Magento 2. Use when you need to decide whether to migrate a M1 module, find alternatives, or skip it.

Creates custom slash commands for Claude Code with proper syntax, frontmatter, arguments, bash execution, and file references. Use when building slash commands, creating custom Claude Code commands, setting up team workflows, or when users mention slash commands, command files, or .md command creation.

Isolated git worktree workflow for coding agents. Handles worktree creation, atomic commits, branch publishing, optional PR creation, and cleanup. Use when implementing code changes that need branch isolation. Trigger terms - worktree, isolated workspace, branch isolation, atomic commits, code changes in isolation, parallel agent work.

Build comprehensive randomization lists for creative entropy. Use when you need to create or expand lists of story elements (professions, locations, objects, names, etc.) for use with entropy tools. Leverages research sources like Kiwix/Wikipedia to build lists with good variety and size.