{"id":"help","name":"help","summary":"インタラクティブなワークスペース発見 - 利用可能なツール、ワークフロー、エージェント、フックについて学ぶ","body":"# /help - Workspace Discovery\n\nGuide users through the capabilities of this workspace setup.\n\n## Usage\n\n```\n/help                    # Interactive guided discovery\n/help workflows          # Workflow orchestration skills\n/help agents             # Specialist agents catalog\n/help tools              # CLI tools (tldr, prove, recall)\n/help hooks              # Active hooks and what they do\n/help advanced           # MCP, frontmatter, customization\n/help <name>             # Deep dive on specific skill/agent\n```\n\n## Behavior Based on Arguments\n\n### No Arguments: Interactive Discovery\n\nUse AskUserQuestion to guide the user:\n\n```\nquestion: \"What are you trying to do?\"\nheader: \"Goal\"\noptions:\n  - label: \"Explore/understand a codebase\"\n    description: \"Find patterns, architecture, conventions\"\n  - label: \"Fix a bug\"\n    description: \"Investigate, diagnose, implement fix\"\n  - label: \"Build a feature\"\n    description: \"Plan, implement, test new functionality\"\n  - label: \"Prove something mathematically\"\n    description: \"Formal verification with Lean 4\"\n```\n\nBased on response, show relevant tools:\n\n| Goal | Show |\n|------|------|\n| Explore codebase | scout agent, tldr CLI, /explore workflow |\n| Fix a bug | /fix workflow, sleuth agent, debug-agent |\n| Build feature | /build workflow, architect agent, kraken agent |\n| Prove math | /prove skill, lean4 skill, Godel-Prover |\n| Research docs | oracle agent, nia-docs, perplexity |\n| Configure workspace | hooks, rules, settings, frontmatter |\n\n### /help workflows\n\nDisplay workflow meta-skills:\n\n```markdown\n## Workflow Skills\n\nOrchestrate multi-agent pipelines for complex tasks.\n\n| Workflow | Purpose | Agents Used |\n|----------|---------|-------------|\n| /fix | Bug investigation → diagnosis → implementation | sleuth → kraken → arbiter |\n| /build | Feature planning → implementation → testing | architect → kraken → arbiter |\n| /debug | Deep investigation of issues | debug-agent, sleuth |\n| /tdd | Test-driven development cycle | arbiter → kraken → arbiter |\n| /refactor | Code transformation with safety | phoenix → kraken → judge |\n| /review | Code review and feedback | critic, judge |\n| /security | Vulnerability analysis | aegis |\n| /explore | Codebase discovery | scout |\n| /test | Test execution and validation | arbiter, atlas |\n| /release | Version bumps, changelog | herald |\n| /migrate | Framework/infrastructure changes | pioneer, phoenix |\n\n**Usage**: Just describe your goal. Claude routes to the right workflow.\n```\n\n### /help agents\n\nDisplay agent catalog:\n\n```markdown\n## Specialist Agents\n\nSpawn via Task tool with subagent_type.\n\n### Exploration & Research\n| Agent | Purpose | Model |\n|-------|---------|-------|\n| scout | Codebase exploration, pattern finding | sonnet |\n| oracle | External research (web, docs, APIs) | sonnet |\n| pathfinder | External repository analysis | sonnet |\n\n### Planning & Architecture\n| Agent | Purpose | Model |\n|-------|---------|-------|\n| architect | Feature planning, design docs | sonnet |\n| plan-agent | Create implementation plans | sonnet |\n| phoenix | Refactoring & migration planning | sonnet |\n\n### Implementation\n| Agent | Purpose | Model |\n|-------|---------|-------|\n| kraken | TDD implementation, refactoring | sonnet |\n| spark | Quick fixes, lightweight changes | haiku |\n\n### Review & Validation\n| Agent | Purpose | Model |\n|-------|---------|-------|\n| arbiter | Test execution, validation | sonnet |\n| critic | Code review | sonnet |\n| judge | Refactoring review | sonnet |\n\n### Investigation\n| Agent | Purpose | Model |\n|-------|---------|-------|\n| sleuth | Bug investigation, root cause | sonnet |\n| debug-agent | Issue investigation with logs | sonnet |\n| profiler | Performance, race conditions | sonnet |\n\n### Documentation & Handoff\n| Agent | Purpose | Model |\n|-------|---------|-------|\n| scribe | Documentation, session summaries | sonnet |\n| chronicler | Session analysis, learning extraction | sonnet |\n```\n\n### /help tools\n\nDisplay CLI tools and capabilities:\n\n```markdown\n## Built-in Tools\n\n### TLDR Code Analysis\nToken-efficient code exploration (95% savings vs reading raw files).\n\n```bash\ntldr tree src/              # File tree\ntldr structure src/ --lang python  # Code structure (codemaps)\ntldr search \"pattern\" src/  # Search files\ntldr cfg file.py func       # Control flow graph\ntldr dfg file.py func       # Data flow graph\ntldr impact func src/       # Reverse call graph (who calls this?)\ntldr dead src/              # Find dead code\ntldr arch src/              # Detect architectural layers\n```\n\n### /prove - Formal Verification\nMachine-verified proofs without learning Lean syntax.\n\n```\n/prove every group homomorphism preserves identity\n/prove continuous functions on compact sets are uniformly continuous\n```\n\nRequires: LM Studio running Godel-Prover model locally.\n\n### Memory System\nStore and recall learnings across sessions.\n\n```bash\n# Recall past learnings\n(cd opc && uv run python scripts/core/recall_learnings.py --query \"hook patterns\")\n\n# Store new learning (via /remember skill)\n/remember \"Hook X works by...\"\n```\n\n### Premortem Risk Analysis\nIdentify failure modes before they occur.\n\n```\n/premortem [plan-file]     # Analyze implementation plan for risks\n```\n```\n\n### /help hooks\n\nDisplay active hooks:\n\n```markdown\n## Active Hooks\n\nHooks extend Claude's behavior at key lifecycle points.\n\n### Session Lifecycle\n| Hook | Event | Purpose |\n|------|-------|---------|\n| session-register | SessionStart | Register session in coordination DB |\n| session-start-recall | SessionStart | Auto-inject relevant learnings |\n| session-end-cleanup | SessionEnd | Cleanup temp files |\n| session-outcome | SessionEnd | Prompt for session outcome |\n\n### User Prompt Processing\n| Hook | Event | Purpose |\n|------|-------|---------|\n| skill-activation-prompt | UserPromptSubmit | Suggest relevant skills |\n| premortem-suggest | UserPromptSubmit | Suggest risk analysis for implementations |\n\n### Tool Interception\n| Hook | Event | Purpose |\n|------|-------|---------|\n| tldr-read-enforcer | PreToolUse:Read | Suggest tldr for large files |\n| smart-search-router | PreToolUse:Grep | Route to ast-grep for structural search |\n| file-claims | PreToolUse:Edit | Track which sessions edit which files |\n| signature-helper | PreToolUse:Edit | Inject function signatures |\n| import-validator | PostToolUse:Edit | Validate imports after edits |\n\n### Validation\n| Hook | Event | Purpose |\n|------|-------|---------|\n| typescript-preflight | PreToolUse:Bash | Type-check before running |\n| compiler-in-the-loop | Stop | Run Lean compiler for /prove |\n\n### Subagent Coordination\n| Hook | Event | Purpose |\n|------|-------|---------|\n| subagent-start | SubagentStart | Initialize agent context |\n| subagent-stop | SubagentStop | Extract learnings from agents |\n```\n\n### /help advanced\n\nDisplay advanced customization:\n\n```markdown\n## Advanced: Customization & Extension\n\n### Skill Frontmatter\nSkills use YAML frontmatter for metadata and tool restrictions:\n\n```yaml\n---\nname: my-skill\ndescription: What it does\ntriggers: [\"keyword1\", \"keyword2\"]\nallowed-tools: [Bash, Read, Edit]\npriority: high\nskills: [other-skill]  # Auto-load dependencies\n---\n```\n\n### Agent Frontmatter\nAgents declare their capabilities:\n\n```yaml\n---\nname: my-agent\ndescription: Specialist for X\nmodel: sonnet|haiku|claude-opus-4-5-20251101\ntools: [Read, Grep, Glob, Bash]\n---\n```\n\n### MCP Servers\nExternal tool integrations:\n\n| Server | Purpose |\n|--------|---------|\n| ast-grep | Structural code search/refactoring |\n| firecrawl | Web scraping |\n| github-search | Search GitHub code/issues |\n| morph | Fast file editing (10k tokens/sec) |\n| nia | Documentation search |\n| perplexity | AI-powered web research |\n\n### Rules (.claude/rules/*.md)\nAlways-on instructions injected into context:\n\n- `claim-verification.md` - Verify before asserting\n- `dynamic-recall.md` - How to use memory system\n- `tldr-cli.md` - TLDR usage reference\n\n### Creating Custom Skills\n```bash\nmkdir -p .claude/skills/my-skill\n# Create .claude/skills/my-skill/SKILL.md with frontmatter\n```\n\n### Creating Custom Agents\n```bash\n# Create .claude/agents/my-agent.md with frontmatter\n```\n```\n\n### /help <specific-name>\n\nFor specific skill/agent lookup:\n\n1. Search for matching skill: `.claude/skills/<name>/SKILL.md`\n2. Search for matching agent: `.claude/agents/<name>.md`\n3. Read the file and display:\n   - Description\n   - Usage examples\n   - When to use\n   - Related skills/agents\n\n## Dynamic Discovery\n\nTo keep help current, dynamically read frontmatter:\n\n```bash\n# List all skills with descriptions\nfor f in .claude/skills/*/SKILL.md; do\n  head -10 \"$f\" | grep -E \"^(name|description):\"\ndone\n\n# List all agents\nfor f in .claude/agents/*.md; do\n  head -10 \"$f\" | grep -E \"^(name|description):\"\ndone\n```\n\n## Example Interaction\n\n**User**: `/help`\n\n**Claude**:\n```\nWhat are you trying to do?\n\n[ ] Explore/understand a codebase\n    Find patterns, architecture, conventions\n\n[ ] Fix a bug\n    Investigate, diagnose, implement fix\n\n[ ] Build a feature\n    Plan, implement, test new functionality\n\n[ ] Prove something mathematically\n    Formal verification with Lean 4\n\n[Other: describe your goal]\n```\n\n**User selects**: \"Fix a bug\"\n\n**Claude**:\n```\n## Bug Fixing Tools\n\n### Quick Path\nJust describe the bug. I'll use the /fix workflow automatically.\n\n### Manual Control\n- **sleuth** agent: Deep investigation, root cause analysis\n- **debug-agent**: Log analysis, state inspection\n- **kraken** agent: TDD implementation of fix\n\n### Commands\n- `/fix` - Full workflow (investigate → diagnose → implement → test)\n- `/debug` - Investigation only\n\n### Tips\n- Provide error messages, logs, reproduction steps\n- The premortem hook will suggest risk analysis before implementation\n```","author":"@parcadei","ownerProfile":null,"authorContacts":null,"sourceUrl":"https://github.com/parcadei/Continuous-Claude-v3/tree/main/.claude/skills/help","license":"MIT","category":"document","lang":"en","tokens":2313,"stars":0,"calls30d":2,"claimed":false,"visibility":"public","origin":"crawler","version":"0.1.0","createdAt":"2026-08-22","updatedAt":"2026-08-22","files":[],"requires":{"mcp":[],"tools":["AskUserQuestion","Bash","Read","Glob","Grep"]},"safety":{"flags":[],"scannedAt":"2026-08-22","hasScripts":false,"networkEndpoints":[]}}