skill-builder
Skill0 starsChanged 50 days agoMITView on GitHub
Interview someone about a task they repeat and turn it into a skill spec ready to build. Trigger when someone says 'I want to automate something', 'build me a skill', 'I keep doing this manually', 'can Claude help with this', 'skill intake', 'skill builder', or describes a tedious repeated task and sounds tired of it.
75 other projects use a version of this.
Adopted by
Most recently taken up by
- SamuelBostic29/claude-skillstheir own version, first seen Nov 16, 2026
- BloodSweatxED/Claude-skillstheir own version, first seen Nov 16, 2026
- arctan303/creator.skilltheir own version, first seen Nov 12, 2026
- Kikobazz123/lordgen-skill-buildertheir own version, first seen Nov 8, 2026
- AsiaOstrich/EngramGraphtheir own version, first seen Nov 8, 2026
- shetengteng/skillix-hubtheir own version, first seen Nov 8, 2026
Used with
Projects with it also use contract-test, literature-research and 10 more.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
squall-chuaSkillcontract-test
Test a running API against its own contract — OpenAPI, GraphQL, protobuf, or AsyncAPI — from the outside, and report every drift between what the contract promises and what the API does.
Also in 3 projects that use this
nickjwellsSkillliterature-research
Search academic literature and perform deep research reviews. Use when user asks to search PubMed, find academic papers, or do literature reviews.
Also in 3 projects that use this
nopandererSkillcommit-and-push
Automatically analyze git changes, generate Conventional Commits compliant commit messages, commit changes, and push to remote in a single automated workflow. Use when the user invokes /commit-and-push or explicitly requests to automatically commit and push changes following Conventional Commits format (feat:, fix:, docs:, etc.). This skill provides a fully automated git workflow without manual intervention.
Also in 3 projects that use this
MicrosoftSkillskill-reviewer
Review skill PRs with structured severity-rated feedback covering token budgets, routing conflicts, and repo conventions. WHEN: \"review skill\", \"review skill PR\", \"review skill changes\", \"check skill quality\", \"skill PR feedback\".
Also in 3 projects that use this
MicrosoftSkilldraft-pr
Creates a pull request for the current branch, generating a clear title and a description that summarizes the changes and includes instructions on how to test them. Use when asked to open, draft, or create a PR.
Also in 3 projects that use this
SkillPilotCHSkilllevel-up
Eine Automatisierung pro Woche finden, scopen, shippen. Output landet als proposed-Eintrag in cadence.md mit konkreten Implementations-Schritten.
Also in 3 projects that use this
- +1 more from microsoft
Other versions75 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 75
- yusufkaraaslan/Skill_Seekers★ 15ksame name
- block/berd★ 1.1ksame name
- boshu2/agentops★ 454same name
- metaskills/skill-builder★ 112same name
- AsiaOstrich/universal-dev-standards★ 84same name
- nateherkai/a-bunch-of-skills★ 41same name
- clasen/Skills★ 40same name
- sanqi-cd/Sanqi-Skills★ 28same name
- arctan303/creator.skill★ 19same name
- aiagentwithdhruv/skills★ 16same name
65 more not listed
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.
- brand-guardrailsSkill
- campaign-copySkill
- design-briefSkill
- moodboard-queriesSkill
- product-descriptionsSkill
- rate-and-improveSkill
- social-trend-scoutSkill
The file
We split this skill into the rules it gives an agent
6 rules. 0 are common, 5 are distinctive, and 1 is very specific to this project.
0
Common
Most of the 66 edited versions we read keep these.
5
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
- Distinctive: line 29
Do not leave the interview until all six are filled.
- Distinctive: line 45
The task: Start broad, then ask "Walk me through the last time you did this.
- Distinctive: line 58
If they share a file, read it and note format, columns, and structure.
- Distinctive: line 62
When the checklist is full, give a 4 to 6 sentence summary and ask "Does that sound right?
- Very specific: line 70
Write to INTAKE_DIR/[skill-name]-[YYYY-MM-DD].md.
- Distinctive: line 86
Never send the spec to anyone.
“Common” is measured: kept by at least 6 in 10 of the 66 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
Skill Builder
A friendly intake interview that captures one repeated task in plain language and produces a structured spec a builder can work from. It asks one question at a time, probes every data source, and saves the spec locally. Nothing gets sent anywhere.
Setup
[FILL IN]
INTAKE_DIR = [folder where specs are saved, e.g. skills/intakes/]
NAME_PREFIX = [none | a namespace you want on every skill. Default: none, use exactly what the person picks]
BUILD_TIERS = [keep the default tiers below or replace with your own]Inputs
- A person with a task they do over and over
- Any files they can share: templates, exports, screenshots, examples of the finished output
Workflow
Completion checklist (track silently)
Do not leave the interview until all six are filled. Re-check after every answer.
- WHO: name, role, team
- WHAT: the task in plain words
- TOOLS: at least one tool, platform, or data source, probed
- STEPS: a walkthrough of at least 3 steps
- OUTPUT: what gets produced and where it goes
- FREQUENCY: how often, and what kicks it off
1. Interview
Open with something like: "Tell me about a task you keep doing over and over that you wish you didn't have to. Describe it like you'd explain it to a friend."
Cover these areas, adapting to what they say:
- Who they are. Role, team, how long they've done this task.
- The task. Start broad, then ask "Walk me through the last time you did this. What did you do first?"
- Tools and data. "What do you have open on your screen when you're doing this?" Run the Data Source Probe on every source mentioned.
- Step by step. "What's the very first thing you do?" then "then what?" until the end. Don't let them skip steps.
- Output. What they end up with, where it goes, who sees it. Ask for an example.
- Pain. What takes longest, what goes wrong most, what they'd make disappear.
- Tricky bits. Judgment calls, exceptions, logins only they have.
Data Source Probe (every file, sheet, report, platform, or feed):
- Where does it come from? Who creates it?
- Same format every time, or does it change?
- Downloaded, copied, or accessed online?
- Can you share an example right now?
If they share a file, read it and note format, columns, and structure. If they can't, ask what a typical row looks like and mark it "described, not provided."
2. Confirm
When the checklist is full, give a 4 to 6 sentence summary and ask "Does that sound right? Anything I missed?" Then ask what to call the skill (short slug, lowercase with dashes). Use exactly what they pick. No more interview questions after this.
3. Write the spec
Fill the template in Output format. Show it and ask for corrections before saving.
4. Save
Write to INTAKE_DIR/[skill-name]-[YYYY-MM-DD].md. Copy any shared files into the same folder and list them under ATTACHED FILES.
5. Wrap up
Tell them where the spec lives and that they can run the intake again to save a new version.
Rules
- One question at a time. Never stack questions.
- Vague answer? Ask for a specific recent example.
- Unknown jargon or tool name? Ask what it actually does for them.
- Tangent: acknowledge briefly, note it if relevant, steer back to the next unchecked item.
- Stalled after about 15 exchanges: name the missing checklist items and ask to cover them quickly.
- Several tasks mentioned: finish the most repetitive one, then offer a second intake.
- Very vague start ("help with marketing stuff"): ask "What did you do today that you didn't enjoy?"
- Anything needing API keys or platform credentials gets flagged prominently in Build Notes as infrastructure work.
- Never send the spec to anyone. Save locally only.
Output format
===== SKILL INTAKE SPEC =====
Generated: [date]
Submitted by: [name], [role], [team]
SKILL NAME: [slug]
TASK SUMMARY: [2-3 plain sentences]
TRIGGER PHRASES: [what someone would say to start it]
TOOLS / SYSTEMS:
- [tool] - [its role in the task]
DATA SOURCES:
- [source]
- Origin:
- Format:
- Access method:
- Sample provided: [Yes, see attachment | No, described as ...]
CURRENT MANUAL STEPS:
1.
2.
3.
DESIRED OUTCOME: [what is produced, where it goes, who sees it]
FREQUENCY: [how often, what triggers it]
EDGE CASES / TRICKY BITS:
CREDENTIALS / ACCESS NEEDED:
ATTACHED FILES:
- [filename] - [what it is]
---
BUILD NOTES
Complexity tier:
1 - single tool, no connectors, text or document work
2 - one or two connectors, simple retrieve then output
3 - three or more connectors, multi-phase, scheduled
Engineering - needs custom code, API setup, or credential deployment
Flags: [unclear items, tools with no connector, missing access, odd data formats]
Buildable as-is: [Yes | Needs clarification | Needs infrastructure first]
===== END SPEC =====Example summary line: "Northwind Apparel's social lead copies weekly post stats from two platforms into a sheet every Monday, then writes a three-bullet recap for the team chat. Takes about an hour."
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.

