Discover

After interview (5-8 exchanges), produce:

• Research specification → quality_reports/research_spec_[topic].md
• Domain profile → .claude/rules/domain-profile.md (if still template)

/discover lit [topic] — Literature Review

Search and synthesize academic literature.

Agents: Librarian (collector) → librarian-critic (reviewer) Output: Annotated bibliography + BibTeX entries + frontier map

Workflow:

1. Read domain-profile.md for field journals and seminal references
2. Dispatch Librarian to search: top-5 journals, field journals, NBER/SSRN, citation chains
3. Dispatch librarian-critic to check coverage, gaps, recency, scope
4. If gaps found, re-dispatch Librarian for targeted search (max 1 round)
5. Save to quality_reports/lit_review_[topic].md

/discover data [requirements] — Data Discovery

Find and assess datasets for the research question.

Agents: Explorer (finder) → explorer-critic (assessor) Output: Ranked data sources with feasibility grades

Workflow:

1. Read research spec and strategy memo if they exist
2. Dispatch Explorer to search: public microdata, admin data, surveys, international, novel sources
3. Dispatch explorer-critic to check measurement validity, sample selection, external validity, ID compatibility
4. Save to quality_reports/data_exploration_[topic].md

/discover ideate [topic] — Research Ideation

Generate structured research questions and hypotheses from a topic or dataset.

Agents: Direct generation (no agent dispatch) Output: Research questions with empirical strategies

Generate:

1. 3-5 research questions with clear hypotheses
2. For each: potential identification strategy, data requirements, expected contribution
3. Rank by feasibility and novelty
4. Save to quality_reports/research_ideas_[topic].md

Principles

• Interview style: Be curious, not prescriptive. Draw out the researcher's thinking.
• Literature honesty: Never fabricate citations. Mark unverified as % UNVERIFIED.
• Data feasibility matters: A perfect dataset you can't access is useless.
• Domain-profile aware: Always read domain-profile.md first for field calibration.
• Worker-critic pairing: Librarian + librarian-critic, Explorer + explorer-critic. Never skip the critic.