research-router-skill — Codex skill: what it does, install
This skill routes broad life-sciences questions to the right skills, normalizes key entities, and produces a concise evidence-backed synthesis. Use it when one lookup is not enough.
npx skills add openai/plugins --skill research-router-skill -a codexWhat research-router-skill does
research-router-skill turns an open-ended research request into a small retrieval plan. It identifies the goal, normalizes genes, diseases, variants, compounds, or datasets, then selects up to three research lanes. For variant work it can route to skills such as clinvar-variation-skill, gnomad-graphql-skill, and ensembl-skill; structure questions can go to alphafold-skill, rcsb-pdb-skill, and uniprot-skill.
The skill gathers evidence, reconciles the results, and structures the response around the user’s question. Its output covers a working conclusion, evidence by lane, caveats, and possible follow-up analyses. When Codex subagents are available, they are reserved for independent evidence lanes, while the coordinating agent handles entity normalization, conflicts, and final synthesis.
This is not the right layer for a narrow single-source lookup already covered by a specific skill. Check that the downstream skills you need are available, because the router selects them but does not replace their retrieval work.
What it helps with
- break a broad life-sciences question into research lanes
- normalize genes, diseases, variants, compounds, or dataset identifiers
- select skills for genetics, structure, chemistry, or clinical evidence
- combine independent subagent findings into one answer
- separate supported conclusions from suggestive or conflicting evidence
Example prompts for the agent
- Research what is known about the link between `BRCA1` and this disease, choose the relevant skills, and synthesize the findings.
- Compare this variant with its phenotype and clinical evidence, then state the cohort and study-design caveats.
- Review this protein and its ligands across independent evidence lanes, then give me a concise synthesis.
How to install research-router-skill
1Skills CLI (Vercel)
npx skills add openai/plugins --skill research-router-skill -a codex2Manually: put the files in place
.agents/skills/research-router-skill/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add openai/plugins --skill research-router-skill -a claude-code2Manually: put the files in place
~/.claude/skills/research-router-skill/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/research-router-skill/
Manually: put the files in place
.cursor/skills/research-router-skill/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install openai/plugins research-router-skill2Manually: put the files in place
.github/skills/research-router-skill/SKILL.mdPlatform CLI
gemini skills install https://github.com/openai/plugins --path plugins/life-science-research/skills/research-router-skillManually: put the files in place
.devin/skills/research-router-skill/SKILL.mdThe legacy .windsurf/skills/ path still works
Claude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
From the same set201
The whole set →Similar
Questions about research-router-skill
How do I install research-router-skill in Codex?
Run npx skills add openai/plugins --skill research-router-skill, then run npx skills add openai/plugins --skill research-router-skill -a codex. The installed skill file is .agents/skills/research-router-skill/SKILL.md.
What does research-router-skill do?
It turns a broad life-sciences question into a retrieval plan, normalizes the main entities, selects downstream skills, gathers evidence, and writes a concise synthesis with caveats.
When does research-router-skill use subagents?
It uses Codex subagents when independent lanes can be researched in parallel, such as genetics, structure, chemistry, and clinical evidence. The coordinating agent keeps entity normalization, conflict resolution, and final synthesis.
