research-swarm
Skill0 starsChanged 3 months agoMITView on GitHub
Use when a question deserves a specialist's answer rather than a quick reply, sizing a market, grounding a claim in science, scouting competitors, a destination, a trend, a regulation. Triggers on "research", "deep dive", "become an expert on", "what's the landscape", or any decision that will spend real money or weeks on the answer. Fans out parallel agents across distinct lenses, verifies the load-bearing claims, and returns a cited, decision-ready brief.
2 other projects use a version of this.
Adopted by
Most recently taken up by
- merttcetn/research-swarmtheir own version, first seen Oct 11, 2026
- tonydzi/second-brain-starter-kittheir own version, first seen Sep 8, 2026
Other versions2 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 2
- tonydzi/second-brain-starter-kit★ 6same name
- merttcetn/research-swarm★ 0same 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 repository5 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- llm-actionsSkill
- product-onboardingSkill
- product-specSkill
- ship-gateSkill
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 8
WHEN a question deserves a specialist, run a swarm, not a single search.
- Distinctive: line 24
Verify before believing: pick the claims the decision actually rests on and send a fresh agent per claim to break it (find …
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
research-swarm
WHEN a question deserves a specialist, run a swarm, not a single search. Scale to stakes: a curiosity gets one agent or none; a decision that spends money or weeks gets the full ladder.
- Pin the question to a decision first: what will this answer decide, by when, and what finding would change the decision. Research with no decision attached is entertainment; say so and ask.
- Decompose into 3 to 6 lenses that see DIFFERENT things, not copies of the same search. The usual spread: the science (what is actually true), the market (who pays, how much, who else sells), the customer (what they experience and say in their own words), the rules (regulation, and which sustainability or health claims you are allowed to make), the field (who does this best today and how).
- Fan out one subagent per lens, all in parallel, each with a written brief: the question, its lens, and the requirement to search the live web, read the actual sources (not result snippets), and return every claim with its source URL and date. An uncited claim is a rumor.
- Verify before believing: pick the claims the decision actually rests on and send a fresh agent per claim to break it (find counter-evidence, trace the number to its original source, check it is current). A claim that survives an attempt to kill it is evidence; one that was merely repeated is not.
- Synthesize one brief, answer first: the recommendation, then what is solid (cited), what is contested (both sides cited), and what is thin. Numbers carry dates and sources. Written for the person deciding, not the person who researched.
- Name what was not covered (a lens skipped, a source paywalled, a claim unverified) instead of letting silence read as coverage.
Skill updated: 2026-08-15
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.