Produce evidence-rich research memos and decision briefs with cited sources, verified claims, and explicit limitations — single-agent-first, fanning out only when parallel work clearly helps.
---
name: deep-research
description: Evidence-based research memos, literature reviews, and multi-source decision briefs with citations, trade-offs, and a clear conclusion. Use for synthesis across many sources; not for quick lookups or single-source summaries.
---
A disciplined research workflow (V8) that separates evidence from analysis and keeps the run diagnosable.
## Workflow
1. Scope first: write a `research-plan.md` and pick the lightest artifact that satisfies the task (brief memo, review, or full report).
2. Go single-agent first; only fan out to parallel sub-agents when that clearly helps. Keep the harness thin and the judgment in the skill.
3. Gather grounded notes into `workspace/research-notes/*.md` and track every source in a `registry.md`, verifying citations rather than trusting recall.
4. Synthesize a `draft.md` that explicitly separates evidence from analysis, states limitations, and ends with a clear conclusion; record an `evaluation.md` and `run-summary.json` for observability.
Full skill & source: https://github.com/staruhub/ClaudeSkills/tree/9ed9d5c2d1ded8d2b401bf3eac09168d62f44bbd/skills/Geek-skills-deep-researchSign in to rate and review this skill.
No reviews yet. Be the first to review this skill.