Next Steer

Wake objective

1. Re-read AGENTS.md, this file, loop-state.md, and any workspace files changed during sleep.
2. Diff operator revisions against the wave-1 artifacts; preserve evidence anchors and fold corrections in.
3. Run the next-wave literature sweep across the six mapped clusters:
   • persistent wiki / compiled markdown knowledge
   • agent memory / episodic continuity
   • repo docs / knowledge graph / wiki maintenance
   • multi-agent orchestration + verification
   • QMD / semantic retrieval / inquiry routing
   • human-in-the-loop documentation governance
4. Convert wave-1 proposals into implementation-ready matrices:
   • Miadi kit: candidate plugins/, agents/, skills/
   • llms-txt: candidate rispecs/ corpus + lint/promotion loop
   • suite routing: QMD / episodic / inquiry-routing handoff model
5. If new operator edits expose missing contract language, patch AGENTS.md before drafting wave 2 outputs.

Expected wave-2 outputs

• literature gap-fill with real paper / source shortlist
• a concrete kit-vs-suite file/folder proposal matrix
• a rispec/wiki automation loop draft
• blocker artifact only if QMD / routing semantics stay too ambiguous for grounded next steps