
Orchestrates the full four-phase complexity audit pipeline for any target path or glob pattern in the Orchestra codebase. Use this skill when a user requests a complexity audit, tech-debt audit, refactor assessment, or complexity analysis. Triggers on: complexity audit, run audit, tech debt audit, audit backend, audit frontend, audit services, audit components, measure complexity, complexity report, refactor audit, code quality audit. Accepts a target path (e.g., `backend/src/services/`, `frontend/src/components/`, or `all`) and an optional existing GitHub issue URL. Handles all setup — issue creation, branch, worktree, .audit/ directory — then delegates execution to the complexity-auditor agent, and finishes with PR reporting.
Explains code with visual diagrams and analogies. Use when explaining how code works, teaching about a codebase, or when the user asks "how does this work?"
Generate a Product Requirements Document (PRD) for a new feature. Use when planning a feature, starting a new project, or when asked to create a PRD. Triggers on: create a prd, write prd for, plan this feature, requirements for, spec out.
Recursive Language Model pattern for processing large inputs using Sonnet orchestrator and parallel Haiku subagents. Decomposes complex tasks into chunks, processes them in parallel via sub-agents, then synthesizes results. Use when: analyze large codebase, recursive analysis, deep analysis, process large input, comprehensive review, rlm, recursive reasoning, review massive document, analyze entire repository.
Template-driven full-stack feature development from user stories. Orchestrates the complete pipeline: issue template reading, duplication check, GitHub issue creation, branch/worktree setup, PRD generation with agent-browser verification, Ralph JSON conversion, commit/push, and Ralph execution. Use when you have user stories ready to implement. Triggers on: feature dev, implement story, build feature from story, story to feature, template-driven, agent-browser verification.
Template-driven bug reproduction and TDD-first fixing pipeline. Orchestrates: bug template reading, duplication check (regression-aware), GitHub issue creation, branch/worktree setup, bug reproduction with agent-browser, root cause analysis, PRD generation with TDD-first story ordering, Ralph JSON conversion, commit/push, and Ralph execution. Use when you have a bug to reproduce and fix. Triggers on: bug report, fix bug, reproduce bug, bug to fix, report bug, tdd fix.
Convert PRDs to prd.json format for the Ralph autonomous agent system. Use when you have an existing PRD and need to convert it to Ralph's JSON format. Triggers on: convert this prd, turn this into ralph format, create prd.json from this, ralph json.
Meta-improvement skill that analyzes conversation context and uncommitted git changes to propose improvements to the AI agent ecosystem itself (agents, skills, commands). Use when the user asks to reflect on the conversation, improve the agent ecosystem, or optimize workflows based on patterns observed. This skill ALWAYS acts as supervisor first, producing an initial analysis report, then spawns specialized builder agents based on user-approved plan. Always requires user approval before making changes.
Run and verify backend tests for the FastAPI/Python application.
Browser automation with persistent page state using Vercel's agent-browser CLI. Use when users ask to navigate websites, fill forms, take screenshots, extract web data, test web apps, or automate browser workflows. Trigger phrases include 'go to [url]', 'click on', 'fill out the form', 'take a screenshot', 'scrape', 'automate', 'test the website', 'log into', or any browser interaction request.
Documentation drift detection between docs.ruska.ai (Docusaurus wiki) and chat.ruska.ai (React+FastAPI app). Provides domain knowledge for drift categories, scoring rubric, crawl strategies, council assessment criteria, and output format templates. Used by the /doc-drift command which orchestrates an Agent Team of crawlers and assessors.
Archive the current Ralph prd.json and progress.txt before starting a new feature. Use when you need to archive a completed or abandoned Ralph run. Triggers on: archive ralph, archive prd, ralph archive, clear ralph, reset ralph.
Run and verify frontend tests for the React/Vite application.
Record browser workflows as video and convert to high-quality GIFs using agent-browser and ffmpeg. Use when users ask to record workflow, create gif, record demo, browser recording, workflow gif, or screen recording.
Apply coordinated changes across multiple files efficiently using parallel Edit tool calls. Use when making systematic changes like renaming variables, updating imports, applying pattern changes, or refactoring across files. Triggers on: rename across files, update all imports, change pattern in multiple files, refactor across codebase, batch edit files.