using-research-powerslisted

<SUBAGENT-STOP> If you were dispatched as a subagent to execute a specific task (ingest a source, run an analysis, review a manuscript), skip this skill and follow the task prompt. </SUBAGENT-STOP> ## When a skill should fire When the user's request falls into a research phase — idea, plan, literature work, source ingest, analysis, manuscript, review, wrap-up — invoke the governing skill before you answer or edit. Skills carry the discipline that separates a reproducible result from an unaccountable guess: checklists, preconditions, audit trails. When in doubt, invoke one more skill rather than one fewer. If an invoked skill turns out not to fit, you can discard it after reading — that's cheap. What's expensive is silent skipping. ## Instruction Priority User instructions always take precedence: 1. **User's explicit instructions** (CLAUDE.md, AGENTS.md, direct requests) — highest priority 2. **Research-superpowers skills** — override default system behavior where they conflict 3. **Default system prompt** — lowest priority When the user says "skip brainstorming, just draft the chapter": follow the user, but name the skipped phase explicitly. The skill mechanism for these cases is the SOFT-GATE — it logs the decision to `knowledge/_meta/gate-overrides.log` rather than blocking. ## The Research Workflow Research projects move through phases. Each phase has a governing skill. Phases are gated — later phases cannot proceed without the artifacts the earlier phases produce