# Research Analyze

> Internal evidence-index analysis and verification stage for the canonical research skill.

- Skill: `runxhq/research-analyze` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds@latest add runxhq/research-analyze`
- Raw SKILL.md: https://api.skillmd.com/api/skills/runxhq/research-analyze/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Research & Search
- Author: runxhq (https://skillmd.com/u/runxhq)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/runxhq/research-analyze

---


# Research Analyze

Analyze only the supplied, runtime-admitted source index. This internal stage is
shared by remote/provider evidence and native local-file evidence so acquisition
cannot fork the research judgment or verification contract.

## Agent task contract

### `research-synthesize`

Preserve the objective exactly and return a bounded `research_candidate` with:

- `decision`: `ready`, `needs_more_evidence`, or `not_worth_publishing`;
- `research_brief`: objective, scope, concise summary, and open questions;
- `evidence_log`: material claims with admitted `source_digest` and matching
  `content_digest`, confidence, and relevance;
- `decision_support`: bounded options whose rationales cite admitted source
  digests;
- `risks`: evidence-bound risks and uncertainty.

Distinguish verified evidence from inference. Do not browse, read other files,
publish, mutate a provider, or invent source identity, delivery, certification,
or settlement claims. Return `needs_more_evidence` when the admitted index does
not support the requested decision.