AI Agent Workflow Collection

Update CLAUDE.md with learnings, preferences, or project context. Use when user says "update claude.md", "add to claude.md", or wants to persist something.

Generate project index for faster codebase navigation. Triggers: 'index codebase', 'create index', 'map project'.

Use when user wants to visualize or present shipkit content as HTML. Triggers: 'create presentation', 'visualize this', 'HTML report'.

Scaffold a tiered design system — principles, tokens, aesthetic direction from project context. Triggers: 'design system', 'design tokens', 'brand guidelines', 'visual direction', 'design principles'.

Define the technical approach — mechanisms, design decisions, stack direction, and component structure. Triggers: 'engineering approach', 'how to build this', 'technical design', 'mechanisms'.

Derive technical performance criteria from the engineering blueprint. Writes goals/engineering.json with response times, reliability, coverage thresholds. Evaluate mode checks actuals against targets.

Process user feedback into investigated bug specs with root cause analysis. Triggers: 'triage feedback', 'process bug reports', 'user testing feedback'.

Discover and install MCP servers for enhanced Claude…

Discover and install skills and plugins from the Claude Code…

Master orchestrator — dispatches direction, planning, and shipping loops sequentially. Auto-loaded via session-start hook.

Capture current metric values for goal evaluation. Triggers: 'log metrics', 'update metrics', 'record measurements'.

Direction loop — defines WHY, WHAT, and strategic foundations. Dispatches producer skills and reviews coherence. Can be invoked standalone or by shipkit-master.

Planning loop — translates definitions into specs, plans, and test cases. Dispatches planning skills and reviews alignment. Can be invoked standalone or by shipkit-master.

Shipping loop — spawns implementation team, verifies quality, gates release. Can be invoked standalone or by shipkit-master.

Use when a spec exists and user wants implementation steps. Triggers: 'how to implement', 'create plan', 'plan this', 'what are the steps'.

Define what to build — features, UX patterns, and differentiators that solve discovered user needs. Triggers: 'product definition', 'what to build', 'features', 'solution design'.

Use when defining users, personas, or user journeys. Triggers: 'who are users', 'create personas', 'user research', 'user stories'.

Derive measurable user-outcome criteria from the product blueprint. Writes goals/product.json (completion rates, UX quality, satisfaction). Evaluate mode compares actuals to targets.

Use when starting a new project or refreshing tech stack context. Triggers: 'scan project', 'what's my stack', 'refresh context', 'generate stack'.

Audit LLM prompt pipeline architecture — decomposition, parallelization, chain integrity, schema validation, fallback paths. Finds structural issues no linter catches.

Visual QA using Playwright as a browser automation library. --setup installs Playwright and creates ui-goals.json; default mode writes inline scripts to navigate, screenshot, and report against goals.

Internal reviewer — assesses strategic artifact coherence. Checks vision/why alignment, goal completeness, stage realism. Dispatched by orch-direction, not for direct use.

Internal reviewer — assesses planning artifact alignment. Checks definitions agree, specs cover roadmap, no gaps. Dispatched by orch-planning, not for direct use.

Review changes across 12 quality dimensions and report findings. Use after a chunk of work or before commit.