research-deep
Skill11 starsChanged 31 days agoMITView on GitHub
Execute web research on specific questions and write findings to Notion `raw/<idea-slug>/` or `raw/shared/` as new raw source pages. Use after `/find-gaps` produces a confirmed plan, or when the user gives a direct research question (e.g. "research X" / "find me data on Y" / "look up Z").
5 other projects use a version of this.
Adopted by
Most recently taken up by
- 200iqlabs/shared-skillstheir own version, first seen Oct 31, 2026
- wrg32786/aigent-ostheir own version, first seen Oct 13, 2026
- garri333/Skillstheir own version, first seen Oct 11, 2026
- Weizhena/Deep-Research-skillstheir own version, first seen Aug 23, 2026
- shreyas-makes/agent-skillstheir own version, first seen Jul 15, 2026
Other versions5 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 5
- Weizhena/Deep-Research-skills★ 2.2ksame name
- wrg32786/aigent-os★ 25same name
- 200iqlabs/shared-skills★ 4same name
- shreyas-makes/agent-skills★ 2same name
- garri333/Skills★ 1same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository7 other files it gives its agents.
- AGENTS.mdAGENTS.md
- find-gapsSkill
- founder-profileSkill
- grill-meSkill
- new-ideaSkill
- onboardingSkill
- wiki-lintSkill
The file
We split this skill into the rules it gives an agent
1 rule. 0 are common, 0 are distinctive, and 1 is very specific to this project.
0
Common
Most of the 4 edited versions we read keep these.
0
Distinctive
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 6
Run targeted web research on one or more research questions and produce raw notes, then propose distilling them into the wiki via …
“Common” is measured: kept by at least 6 in 10 of the 4 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Run targeted web research on one or more research questions and produce raw notes, then propose distilling them into the wiki via docs/agent/workflows.md. Maintaining the wiki is the LLM's job, not the user's — don't leave the raw notes sitting unintegrated.
Identify the question. From a /find-gaps plan or a direct user prompt. If vague, narrow it before searching ("research the market" → "What 2024-2025 sources support or contradict the TAM figure cited in the overview's Market section?"). Read the relevant Notion raw/<idea-slug>/ page and raw/shared/ first to avoid retreading ground — if a recent note already covers the question, say so and ask the user whether to skip, supplement, or refresh.
Plan and execute searches. Draft 2-4 targeted queries per question and run them. Read the most relevant sources end-to-end, not just snippets — snippets lose date and context. For multiple research questions in one session, consider launching one sub-agent per question so they run concurrently; each sub-agent owns its question end-to-end (queries, source reading, raw-note write) and returns when done. The parent agent then compiles the report-back.
Write one raw source page per research question under the relevant Notion raw/<idea-slug>/ page, or under raw/shared/ if it applies across ideas. Title it with a date and topic, e.g. 2026-04-24 market pricing research. Each raw page should:
- State the originating question at the top (one sentence).
- Capture findings with source URL, publication date, and direct quote for every material claim.
- Mark uncertain or contested claims as
[uncertain]with a one-line reason. - Surface contradictions across sources explicitly instead of picking the most flattering one.
- Distinguish facts (what a source says) from synthesis (what the agent concludes from multiple sources).
- End with a
## Suggested wiki integrationsection: which idea overview / deep-dive section would update, and a one-line draft of the diff.
Report back. After all notes are written (or all sub-agents have returned), list them with a one-line summary each, ordered by what would most move the verdict. Flag findings that confirm a load-bearing claim, findings that kill one, and questions where the search returned no useful evidence.
Suggest the ingest pass. Once the raw pages exist, propose running the ingest workflow in docs/agent/workflows.md next. Offer to do it directly. Don't leave new evidence as raw-only.
Cite real URLs only — never fabricate one. If a search returns nothing useful, say so explicitly rather than padding the note with adjacent content.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.