mycelium

You are a **Learning Agent** specializing in knowledge capture, pattern detection, and institutional learning.

You are a **Code Quality Reviewer** specializing in security, performance, architecture, and maintainability analysis.

You are a **Specification Compliance Reviewer** specializing in verifying that implementations match stated requirements.

This skill should be used during Phase 4.5B (Context Sync) when context window usage exceeds 50% (auto-triggers at 80%), after each task completion in long sessions, or when spawning fresh agents. Manages context window efficiently to prevent information loss, enables seamless handoffs between agent instances, and maintains coherence across long-running work.

Captures learnings from completed work to grow the knowledge layer. Use when user says "capture this", "document learnings", "save what we learned", or after finalizing work. Stores solutions, decisions, conventions, preferences, anti-patterns, and effective prompts. Does NOT handle commits, PRs, or pattern detection (see mycelium-finalize and mycelium-patterns).

Asks clarifying questions ONE at a time to understand user requirements, using cached capabilities from context loading. Use when user says "clarify this", "what do you need to know", "ask me questions", or when requirements are ambiguous. Determines if external research (Phase 1.5) is needed before planning.

Loads project context, institutional knowledge, and discovers all available capabilities (skills, agents, MCPs). Use when user says "load context", "discover capabilities", "what skills are available", or at the start of any workflow. Caches discovered capabilities in state.json for use by planning phase.

Resumes interrupted workflow from last checkpoint with context-aware scope detection. Use when user says "continue", "resume work", "keep going", "finish this", or after interruption. Supports --full to run all remaining phases (plan→work→review→capture) and --track to switch between multiple plans. Auto-detects whether to finish current phase or run full workflow.

Creates git commit with Co-Author and pull request to finalize completed work. Use when user says "commit this", "create PR", "finalize changes", or after review passes. Handles platform detection (GitHub/GitLab/Gitea) and uses appropriate CLI tools (gh/glab/tea).

Executes complete autonomous development workflow from planning to knowledge capture. Use when user says "build [feature]", "implement [functionality]", "fix [bug]", "debug [issue]", "investigate [problem]", "optimize [component]", or asks technical questions requiring systematic solution. Handles feature development, debugging, investigation, and optimization with full plan → work → review → capture cycle.

Detects recurring patterns from 3+ similar solutions and promotes them to critical-patterns.md. Use when user says "detect patterns", "find recurring issues", "what patterns exist", or after completing work. Recommends skill generation for patterns occurring 5+ times or critical patterns.

Creates structured implementation plans with TDD task breakdown, dependency management, and capability assignment. Use when user says "plan this", "create a plan for [task]", "break down [feature]", "how should we implement [x]", "list all plans", or "switch to [plan]". Assigns best-fit agent, relevant skills, and model tier to each task using cached capabilities from context loading phase (Phase 0).

Performs comprehensive two-stage code review - spec compliance (blocking gate) followed by parallel quality assessment (security, performance, architecture). Use when user says "review this", "check the code", "is this ready", "review my changes", or after completing implementation. Stage 1 must pass before Stage 2 runs.

Bootstraps new or existing projects with .mycelium/ workflow structure using progressive disclosure. Follows critical interactive rules (one question per turn, suggested answers, max 5 per section). Use when user says "setup mycelium", "initialize project", or when starting a new project. Git is optional but recommended for parallel features.

Displays current workflow state, progress dashboard, and active plans. Use when user says "show status", "what's the progress", "where are we", "list tasks", "show current state", or needs overview of work in progress. Supports --verbose for detailed output including metrics and checkpoints.

Previews complete workflow plan without execution (dry-run mode). Use when user says "show me the plan", "what would this look like", "preview [task]", "dry run", or wants to see task breakdown before committing. Supports --verbose to display workflow diagram. No implementation, tests, or reviews executed.

Executes implementation tasks following strict TDD methodology with evidence-based verification. Use when user says "implement this", "work on [task]", "execute the plan", "build task [id]", or "run all tasks". Enforces RED→GREEN→REFACTOR cycle, shows actual test output, supports parallel execution via worktrees.

This skill should be used when encountering blockers, repeated failures (tests failing 3+ times), stuck states, error conditions, or degraded performance during any workflow phase. Provides systematic recovery procedures, escalation paths, rollback protocols, and debugging guidance to handle failures gracefully without compounding problems.

This skill should be used when the user asks to "implement a feature", "write code", "add functionality", "fix a bug", "refactor code", or during any code implementation phase. Enforces the Iron Law of TDD - tests MUST be written first and all tests must pass before implementation is considered complete. 11-step RED→GREEN→REFACTOR cycle.

This skill should be used during Phase 4.5 (Verification) when the user requests "verify this", "check if it works", "run tests", "prove it works", or after implementation completion. Enforces evidence-based validation - no "should work" claims allowed, requires actual test execution with output shown and proof of correctness.