specialize
Skill1 starsChanged 3 months agoMITView on GitHub
Before assisting on a subject where basic intuition is not enough (a method, psychology, any discipline with a literature of its own), search for theory and evidence, come back with verified competence and crystallise it into a knowledge pill. Use for "specialise on X", "go find the best practices", or when about to set up a path or an analysis on general knowledge alone, in a field where established research exists.
1 other project uses a version of this.
Adopted by
Most recently taken up by
- wellwelwel/lagunetheir own version, first seen Oct 31, 2026
Other versions1 project uses a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 1
- wellwelwel/lagune★ 177same 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.
- template/CLAUDE.mdCLAUDE.md
- knowledge-pillSkill
- librarianSkill
- onboardingSkill
- project-kickoffSkill
- session-logSkill
- system-checkSkill
The file
Files
Specialize
v1.0
Goal: before assisting the user on a subject with a literature of its own, the assistant specialises instead of answering on intuition — it searches, comes back with verified competence, and crystallises it into a pill.
Steps
- Define the gap — list the 3-6 questions that basic knowledge would answer with received wisdom. Those are the ones to verify.
- Check existing assets — read
00_context/knowledge_pills/index_knowledge-pills.md: if a pill already covers the subject, use that one (updating it if needed) rather than creating a duplicate (Rule 4 ofCLAUDE.md). - Targeted search — one search per question, not a generic one: aim for studies with an author and a year, meta-analyses, effect sizes. Prefer primary evidence or evidence syntheses over motivational content; flag it when the evidence contradicts common sense — that is the main value of the exercise.
- Distil into a pill — write it in
00_context/knowledge_pills/, one paragraph per pillar (what the evidence says + an "In practice:" line), a typical-mistakes section, linked sources. Dense, not long. Entry in the index in the same operation (Rule 1). Detailed format: knowledge-pill. - Come back with a position — not a summary of the research: the answer to the concrete case in the light of the evidence, including the corrections to whatever was about to be done.
Hard rules
- Different from knowledge-pill: that one distils competence already surfaced in the sessions, this one goes and fetches it from outside.
- Without access to an external search this skill is not run from memory: state the limit and carry on, declaring that the answer is intuition, not verified.
- Evidence is cited, not invoked: a pill with no linked sources is not the output of this skill.
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.