Interview Command

When the user runs /interview, conduct a structured requirements-gathering interview.

Usage

/interview                           # Interactive - ask what to interview about
/interview <topic>                   # Interview about a specific topic
/interview <file-path>               # Use file as context for interview
/interview <topic> --voice           # Enable TTS for questions (requires claude-mlx-tts)
/interview <topic> --template=<type> # Use specific template (feature, bug, refactor, integration)
/interview --blitz <bead-id>         # Blitz mode: rapid clarification for existing bead

Flags

Flag	Description
--voice	Enable TTS for questions (requires claude-mlx-tts)
--template=<type>	Use specific template (feature, bug, refactor, integration)
--blitz	Blitz mode for existing beads (3-10 questions vs 10-40+)

Modes

Standard Mode (default)

Full interview for new specifications:

1. Parse input: Detect if argument is a topic string or file path
2. Load skill: Use the interview skill from skills/interview/SKILL.md
3. Load context and detect TTS in parallel: <!-- PARALLELIZATION: These are independent operations with no shared state -->
   • a) Load context: If file path provided, read file contents; detect work type from input
   • b) Detect TTS: Invoke tts-status skill (see TTS Detection below)
   • If TTS detection fails, mention it and proceed text-only (graceful degradation)
4. Check beads availability: Look for .beads/ directory
5. Run interview: Conduct adaptive 9-phase interview (10-40+ questions based on scope)
6. Generate spec: Output structured markdown specification
7. Optional: Create beads epic/subtasks if beads is available

Blitz Mode (--blitz)

Rapid clarification for existing beads:

1. Validate bead ID: Check format and existence
2. Load bead context: Use references/bead-context.md for rich context loading
3. Assess completeness: Score bead for gaps and missing info
4. Run blitz interview: Adaptive 3-10 questions based on work type and completeness
5. Update bead: Apply clarifications via bd update
6. Optional: Append to existing spec if one exists

Blitz examples:

/interview --blitz bd-a3f8      # Blitz on root bead
/interview --blitz bd-a3f8.2    # Blitz on subtask

Note: /blitz <bead-id> is a shorthand for /interview --blitz <bead-id>

TTS Detection

CRITICAL: To detect TTS, invoke the skill - do NOT use curl or HTTP requests.

Invoke skill: claude-mlx-tts:tts-status

DO NOT:

• ❌ Use curl to probe HTTP endpoints
• ❌ Guess ports or health check paths
• ❌ Run shell commands to detect TTS

Detection results:

• If skill runs and says "running" → TTS available
• If skill runs and says "not running" → invoke /tts-start, retry
• If skill fails or doesn't exist → inform user: "TTS not available, continuing without voice"

Mode-specific behavior:

Mode	TTS Behavior
Standard	Ask user preference, or use --voice flag
Blitz	Auto-detect TTS, enable if available (no prompt)

⚠️ Visual Output Is Always Required

TTS is SUPPLEMENTARY, not a replacement for visual output.

Every message that is voiced MUST also be printed to the terminal. This applies to:

• Questions
• Progress indicators
• Summaries
• Phase transitions

TTS adds voice ON TOP OF visual output. It never replaces it.

See the interview skill for detailed interview logic and phases.