By name: agentdocs description: Browse Agent Kit documentation directly in terminal. Use for quick access to architecture, skills, workflows, and reference docs. disable-model-invocation: true allowed-tools: Read, Glob, Bash argument-hint: [section]
Agent Docs: Terminal Documentation Browser
Browse the Lucid Labs Agent Kit documentation without leaving the terminal.
Usage
/agentdocs # Show documentation overview
/agentdocs [section] # Jump directly to section
/agentdocs search [query] # Search across all docs
Sections
When called without argument, show this menu:
┌─────────────────────────────────────────────────────────────────────────────────┐
│ │
│ AGENT KIT DOCUMENTATION │
│ │
│ ───────────────────────────────────────────────────────────────────────── │
│ │
│ GETTING STARTED │
│ [1] overview - Was ist das Agent Kit? │
│ [2] quickstart - Schnellstart für neue Projekte │
│ [3] stack - Tech Stack Übersicht │
│ │
│ ARCHITECTURE │
│ [4] architecture - Systemarchitektur & Patterns │
│ [5] aidd - AIDD Workflow (Adaptive AI Discovery & Delivery) │
│ [6] piv - PIV Loop (Plan-Implement-Validate) │
│ │
│ DEVELOPMENT │
│ [7] skills - Alle verfügbaren Skills │
│ [8] mastra - AI Agents & Tools mit Mastra │
│ [9] convex - Realtime Database │
│ [10] design - Design System & UI Patterns │
│ │
│ INTEGRATION │
│ [11] linear - Linear Project Management │
│ [12] mcp - MCP Server Setup │
│ [13] productive - Productive.io Time Tracking │
│ │
│ DEPLOYMENT │
│ [14] deployment - Docker, Elestio, CI/CD │
│ [15] scaling - Skalierung & Performance │
│ │
│ ───────────────────────────────────────────────────────────────────────── │
│ │
│ Wähle [1-15] oder tippe einen Suchbegriff: │
│ > _ │
│ │
└─────────────────────────────────────────────────────────────────────────────────┘
Section Mappings
| Section | File(s) |
|---|---|
overview |
.claude/agent-kit-for-ai.md |
quickstart |
CLAUDE.md (Quick Start section) |
stack |
.claude/agent-kit-stack-overview.md |
architecture |
.claude/reference/architecture.md |
aidd |
.claude/reference/aidd-methodology.md |
piv |
CLAUDE.md (PIV Loop section) |
skills |
.claude/skills/README.md + list all skills |
mastra |
.claude/reference/mastra-best-practices.md, mastra/README.md |
convex |
convex/README.md |
design |
.claude/reference/design-system.md |
linear |
.claude/reference/linear-setup.md |
mcp |
.claude/reference/mcp-servers.md |
productive |
.claude/reference/productive-integration.md |
deployment |
.claude/reference/deployment-best-practices.md |
scaling |
.claude/reference/scaling.md |
Direct Section Access
/docs overview
┌─────────────────────────────────────────────────────────────────────────────────┐
│ │
│ AGENT KIT OVERVIEW │
│ │
│ ───────────────────────────────────────────────────────────────────────── │
│ │
│ Das Agent Kit ist ein modulares Engineering Toolkit für AI-Agent-Projekte. │
│ │
│ KERNKONZEPTE: │
│ │
│ • Upstream/Downstream Model │
│ - Agent Kit = Upstream (Template) │
│ - Dein Projekt = Downstream (Fork) │
│ - /sync holt Updates, /promote teilt Verbesserungen │
│ │
│ • AIDD Workflow │
│ - Exploration → Decision → Delivery → Review │
│ - Linear-basiertes Ticket-Management │
│ │
│ • PIV Loop │
│ - Plan → Implement → Validate │
│ - Keine Phase überspringen! │
│ │
│ ───────────────────────────────────────────────────────────────────────── │
│ │
│ Weiter: [stack] [architecture] [skills] [zurück] │
│ │
└─────────────────────────────────────────────────────────────────────────────────┘
/docs skills
┌─────────────────────────────────────────────────────────────────────────────────┐
│ │
│ VERFÜGBARE SKILLS │
│ │
│ ───────────────────────────────────────────────────────────────────────── │
│ │
│ SESSION │
│ /prime Load project context, start session │
│ /session-end End session, update Linear, save time │
│ │
│ PLANNING │
│ /create-prd Create Product Requirements Document │
│ /plan-feature Plan feature implementation │
│ /init-project Initialize new downstream project │
│ │
│ IMPLEMENTATION │
│ /execute Execute an implementation plan │
│ /commit Create formatted git commit │
│ │
│ VALIDATION │
│ /visual-verify UI verification via agent-browser │
│ /pre-production Security & quality check before deploy │
│ │
│ INTEGRATION │
│ /linear Linear project management │
│ /productizer Bridge Linear ↔ Productive.io │
│ /notion-publish Publish markdown to Notion (private) │
│ /clone-skill Clone skills from central repository │
│ /publish-skill Share your skills with the team │
│ /agentdocs This documentation browser │
│ │
│ MAINTENANCE │
│ /sync Sync updates from upstream │
│ /promote Promote patterns to upstream │
│ │
│ ───────────────────────────────────────────────────────────────────────── │
│ │
│ Detail: /docs skill [name] Beispiel: /docs skill commit │
│ │
└─────────────────────────────────────────────────────────────────────────────────┘
/docs stack
Zeige den Inhalt von .claude/agent-kit-stack-overview.md formatiert an.
/docs skill [name]
# Lade Skill-Definition
cat .claude/skills/[name]/SKILL.md
Zeige formatiert:
- Name & Description
- Allowed Tools
- Usage Examples
- Process Steps
Search
/docs search [query]
# Suche in allen Dokumenten
grep -r -i "[query]" .claude/ --include="*.md" | head -20
Output:
┌─────────────────────────────────────────────────────────────────────────────────┐
│ │
│ SUCHERGEBNISSE: "convex" │
│ │
│ ───────────────────────────────────────────────────────────────────────── │
│ │
│ [1] CLAUDE.md:142 │
│ "Database: Convex - Realtime reactive database" │
│ │
│ [2] .claude/reference/architecture.md:78 │
│ "Convex provides automatic real-time updates to UI" │
│ │
│ [3] .claude/agent-kit-stack-overview.md:45 │
│ "STANDARD: Convex (Realtime, Type-safe, Built-in Vector)" │
│ │
│ ───────────────────────────────────────────────────────────────────────── │
│ │
│ Öffnen: Wähle [1-3] für Details │
│ │
└─────────────────────────────────────────────────────────────────────────────────┘
Navigation
Innerhalb jeder Section:
| Key | Action |
|---|---|
[zurück] |
Zurück zum Hauptmenü |
[weiter] |
Nächste verwandte Section |
[1-n] |
Springe zu Subsection |
q |
Beenden |
Implementation Notes
- Alle Docs im Terminal - Kein Finder, kein Browser
- Formatierte Ausgabe - ASCII-Boxes für Lesbarkeit
- Navigation - Einfache Zahlen/Keywords
- Search - Grep-basiert, schnell
- Context-aware - Zeigt relevante "Weiter"-Links
Examples
# Starte Docs-Browser
/docs
# Direkt zu Skills
/docs skills
# Skill-Details
/docs skill commit
# Suche
/docs search "linear"
# Stack Übersicht
/docs stack