research-deep · wrg32786 — Claude Code skill: features and installation
The research-deep Claude Code skill combines vault notes, web sources, and GitHub research into a cited briefing with evidence and a recommendation. It runs on requests to investigate a topic or compare sources.
MIT ★ 26 updated 2026-11-16
npx skills add wrg32786/aigent-os --skill research-deep -a claude-codeWhat research-deep does
The skill searches vault/concepts/ for relevant notes and reads 3 to 5 of them before forming a working hypothesis. It then checks that hypothesis against web sources and GitHub. Web searches use WebFetch or Tavily when available; GitHub research applies when the topic concerns a tool, library, or framework.
The briefing lays out evidence for and against the hypothesis, a recommendation, research gaps, and a confidence level. Each claim needs a citation pointing to a URL or file path. One source alone is not enough; the process calls for triangulation.
With Write, the skill saves the briefing under vault/concepts/research/ or to a path given by the user. That tool is restricted to the briefing artifact and cannot write other files.
What it helps with
- compare vault notes with web and GitHub sources
- test a working hypothesis against supporting and opposing evidence
- create a cited research briefing with a recommendation
- record research gaps and a confidence level
Example prompts for the agent
- Research this library and write a briefing based on the available sources.
- Compare notes in
vault/concepts/with GitHub research on this topic. - Build a research brief with evidence for and against the hypothesis.
How to install research-deep
1Skills CLI (Vercel)
npx skills add wrg32786/aigent-os --skill research-deep -a claude-code2Manually: put the files in place
~/.claude/skills/research-deep/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/research-deep/
1Skills CLI (Vercel)
npx skills add wrg32786/aigent-os --skill research-deep -a codex2Manually: put the files in place
.agents/skills/research-deep/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add wrg32786/aigent-os --skill research-deep -a cursor2Manually: put the files in place
.cursor/skills/research-deep/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install wrg32786/aigent-os skills/research-deep2Manually: put the files in place
.github/skills/research-deep/SKILL.md1Skills CLI (Vercel)
npx skills add wrg32786/aigent-os --skill research-deep -a gemini-cli2Platform CLI
gemini skills install https://github.com/wrg32786/aigent-os --path skills/research-deepManually: put the files in place
.devin/skills/research-deep/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add wrg32786/aigent-os --skill research-deep -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about research-deep
How do I install research-deep in Claude Code?
Run npx skills add wrg32786/aigent-os --skill research-deep -a claude-code. The skill file is located at ~/.claude/skills/research-deep/SKILL.md.
What does research-deep do?
It examines vault notes, web sources, and GitHub, then writes a briefing with cited evidence, a recommendation, research gaps, and a confidence level.
Why does research-deep need more than one source?
It does not produce a briefing from a single source. The skill compares material from the vault, web, and GitHub, and requires citations for claims.
Source text
SKILL.md · MIT
--- name: research-deep agent: newton description: "Synthesize vault, web, and GitHub research into a cited briefing with evidence and recommendations. Use when investigating a topic or requesting multi-source research." allowed-tools: Read, Grep, Glob, WebFetch, Agent, Write user-invocable: true triggers: - research this - deep research - investigate topic - multi-source research - research and synthesize - build me a briefing - what does the evidence say - research brief - full research on --- # Research Deep You are Newton performing a multi-source deep research pass. You assemble evidence — the composer makes the call. ## What this skill does Go wide across vault, web, and GitHub. Triangulate sources. Synthesize into a structured briefing: hypothesis → evidence for → evidence against → recommendation. Every claim traces to a source. ## Protocol ### Step 1: Vault prior art first Before any web search: - Grep `vault/concepts/` for the topic's key terms - Read the top 3-5 relevant notes found - Note what's already documented — never re-investigate settled questions ### Step 2: Formulate a working hypothesis Based on vault context, state a working hypothesis in one sentence. This is the claim you'll test. ### Step 3: Go wide (parallel fetches) Run simultaneously: - 2-3 web searches (via WebFetch or Tavily if available) on different facets of the topic - GitHub search if the topic involves a tool, library, or framework - Any primary source URLs mentioned in the vault ### Step 4: Synthesize Structure the briefing: ``` ## Research Briefing: <topic> **Working hypothesis:** <one sentence> ### Evidence For - <claim> [Source: <URL or path>] - ... ### Evidence Against - <claim> [Source: <URL or path>] - ... ### Recommendation <What the evidence supports. Stated directly. The composer decides — Newton recommends.> ### Gaps <What couldn't be found or confirmed> ### Confidence: High / Medium / Low <Rationale for confidence level> ``` ### Step 5: Write the briefing artifact Use the Write tool to save the briefing to `vault/concepts/research/` or the path the user specified. Return the path. ### Step 6: Return honesty ledger ``` ## Honesty Ledger **Sources checked:** N **Sources that yielded findings:** N **Gaps:** <what's missing> **Confidence level:** High / Medium / Low **Stopped-short:** <scope boundary or time constraint> ``` ## Constraints - Never return a briefing from a single source. Always triangulate. - Every claim has a citation. No citation = no claim. - Vault prior art first. Do not re-research settled questions. - Run vault reads and web fetches in parallel. - Write tool for the briefing artifact only — no other file writes.
© wrg32786 · MIT · wrg32786/aigent-os
Similar skills
SKILL.md skills catalog →research-deep by other authors comes first.