Skills / Departments / Decisions / Skills / Decision Memo
All skills

/ skill

decision-memo

Departments/Decisions/Skills/decision-memo.md

Run the full dialectic and produce a decision memo for Adam. Optionally grounded in NotebookLM-backed libraries.

Open as raw file

Run this skill

Heads up — this skill doesn’t have a structured form yet, so type the full prompt with all decisions baked in (the subprocess can’t answer follow-up questions). Edit me if needed.

Runs with bypassPermissions — local dev only.

decision-memo

Procedure

  1. Resolve scope-relevant context:
  2. If libraries provided: read each Wiki/Libraries/{slug}.md ## Cached summary for fast context. The libraries become available grounding for steps 3.
  3. Invoke dialectic-cycle (with libraries passed through) to run Steelman → Devils-Advocate → Synthesizer.
  4. Synthesizer writes memo using decision-memo template.
  5. Mirror as wiki node at Wiki/Decisions/{slug}.md with node_kind: decision. If libraries were used, add them to links: and tag #source/notebooklm.
  6. Append Log entry — note which libraries (if any) grounded the memo.

When to ground with libraries

  • Personal high-stakes decisions: consider [adam-consumption] (your full content corpus) plus any topic-specific library.
  • Business decisions: topic-specific libraries (e.g., [leadership], [multi-brand-ops]) plus competitive-research libraries ([competitive-{brand-slug}]) if relevant.
  • Quick ops decisions: skip libraries — overhead exceeds value.

Hard rules

  • Never pass libraries containing sensitive content (Adam's MI synthesis, brand compliance docs) — those wouldn't be in libraries anyway, but the skill should sanity-check.
  • The memo MUST include a "Grounding" section listing which libraries were consulted and which citations from notebooklm-query were load-bearing in the recommendation. Provenance matters for decisions you'll revisit later.