Research-to-Article Pipeline

Operator Context

This skill orchestrates the complete content pipeline: parallel research gathering, structured compilation, voice generation, validation, and output. It implements the Pipeline Architecture pattern — gather, compile, ground, generate, validate — with Domain Intelligence embedded in voice-first content methodology.

Hardcoded Behaviors (Always Apply)

• CLAUDE.md Compliance: Read and follow repository CLAUDE.md before starting pipeline
• Over-Engineering Prevention: Generate the article requested. No speculative additions, no "while I'm here" improvements to other content
• Parallel Research: Launch 5 parallel research agents for comprehensive coverage
• Research-Voice Separation: Research informs but NEVER dominates narrative. Raw data must be transformed to story
• Deterministic Validation: Always validate final output with voice_validator.py
• Wabi-Sabi Authenticity: Natural imperfections are features, not bugs. Do not over-polish
• Artifact Persistence: Save research doc AND article to disk. Context is ephemeral, files persist

Default Behaviors (ON unless disabled)

• 5 Research Agents: Primary domain, narrative arcs, external context, community reaction, business context
• Research Compilation: Structure findings with Story Arc first, key facts second
• Voice Mode Selection: Auto-select mode based on content type (profile, awards, journey, etc.)