Skip to main content

ScribeBot Content Pipeline (Research → Drafts → Distribution)

This runbook documents the lightweight production lane for Valkyr Labs content operations.

Objective​

Convert research and project-board signals into publish-ready artifacts with durable traceability.

Inputs​

  • GitHub Project board items (content/marketing + adjacent product issues)
  • strategic constraints and positioning from the active memory bank
  • ThorAPI model/domain updates (for product-grounded messaging)

Output targets​

Per cycle, ScribeBot should emit:

  • Blog draft(s) in web/typescript/valkyr_labs_com/content/drafts/
  • Social thread drafts in web/typescript/valkyr_labs_com/content/social/
  • Docs updates under docs/docs/Products/ValkyrAI/
  • Changelog entry in the ValkyrAI changelog

Data-grounding rule (ThorAPI-first)​

Content must be anchored in generated model/service reality when discussing product behavior.

Recommended references:

  • @thor/model entities (example: MemoryEntry, ContentData)
  • @thor/redux/services/*Service hooks for practical integration examples

Draft quality checklist​

  • Clear problem statement in first 3 paragraphs
  • One concrete implementation pattern (code or API flow)
  • Traceability to issue/epic IDs when relevant
  • Avoid speculative claims that read as product commitments
  • Include operator-level value (support, product, sales, security)

Distribution checklist​

  • Open PR with all generated assets in one batch
  • Add a concise PR body with linked issue briefs
  • Post throughput update to Discord #general with:
    • brief claimed
    • draft files created
    • PR link
    • blockers (if any)

Notes​

This pipeline is intentionally simple: fewer moving parts, faster output, easier audit.