sketch
Skill0 starsChanged 40 days agoMITView on GitHub
Fast throwaway-friendly implementation of a feature whose shape isn't settled yet — a short brief, code on a sketch branch, run it, iterate on feedback, then merge, harden or drop. No milestone docs, ledger, review or doc-sync. Use when the user runs /sketch <what>, or asks for "a quick spike", "a rough version to try", "prototype this fast".
28 other projects use a version of this.
Adopted by
Most recently taken up by
- fullerzz/agent-skillstheir own version, first seen Nov 8, 2026
- jskswamy/skillstheir own version, first seen Nov 8, 2026
- hemal08ce094/DesignSkillsSwiftUItheir own version, first seen Nov 8, 2026
- ashleysbrain/skillstheir own version, first seen Nov 8, 2026
- lulufoo/lulu-boardtheir own version, first seen Nov 8, 2026
- delorenj/skillstheir own version, first seen Nov 7, 2026
Used with
Projects with it also use touchdesigner-mcp, ant 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.
john-data-chenSkilltouchdesigner-mcp
Control a running TouchDesigner instance via twozero MCP — create operators, set parameters, wire connections, execute Python, build real-time visuals. 36 native tools.
Also in 7 projects that use this
bergsideSkillant
Structured, enterprise-focused design system emphasizing clarity, consistency, and efficiency for data-dense web applications.
Also in 3 projects that use this
bergsideSkillartistic
High-contrast, expressive style with creative typography and bold color choices for visually striking interfaces.
Also in 3 projects that use this
- +9 more from bergside
Other versions28 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 28
- bergside/awesome-design-skills★ 2.9ksame name
- glebis/claude-skills★ 414same name
- buildomator/buildomator★ 102same name
- delorenj/skills★ 17same name
- G1Joshi/Agent-Skills★ 15same name
- lulufoo/lulu-board★ 8same name
- lulufoo/lulu-drawer★ 8same name
- Samin12/skills★ 8same name
- nauvalazhar/skills★ 4same name
- zombico/mojulo★ 4same name
18 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.
- CLAUDE.mdCLAUDE.md
- main-agent/CLAUDE.mdCLAUDE.md
- followupSkill
- implementSkill
- impl-planSkill
- milestoneSkill
- walkthroughSkill
The file
We split this skill into the rules it gives an agent
3 rules. 0 are common, 1 is distinctive, and 2 are very specific to this project.
0
Common
Most of the 27 edited versions we read keep these.
1
Distinctive
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 15
No ceremony: no plan or milestone docs, no follow-up ledger entries, no doc-sync, no README/CLAUDE.md edits.
- Distinctive: line 23
Brief: Ask at most 3 questions, only ones whose answer changes what you build.
- Very specific: line 29
Harden: leave the branch; write a hand-off summary (what was built, files, decisions made along the way, known gaps and shortcuts) and …
“Common” is measured: kept by at least 6 in 10 of the 27 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
/sketch — try a feature fast, decide later
The goal is a running version the user can try within minutes, not a finished feature. Speed over polish; nothing here is planned, reviewed or documented until the user decides to keep it.
Hard rules
- Work on a
sketch/<slug>branch created from the current branch (<slug>: 2–4 words of the request). Uncommitted changes on the current branch → stop and ask first. Never push. - No ceremony: no plan or milestone docs, no follow-up ledger entries, no doc-sync, no README/CLAUDE.md edits. No code review before commits — this is the explicit exception to the global "run the review agent before commits" rule; the review happens when the sketch is hardened.
- Checks are cheap ones only: the build, plus the existing tests of the area you touched, must stay green. Add a new test only when it is quicker than checking by hand. No screenshot suites, packaging or device builds.
- Commit after each working step (
Sketch <slug>: <what>), so any iteration can be undone. - Implement in this session. Only when the user explicitly asks for multiple agents (
--agents, "use agents", "in parallel"): split the work by files into 2–4 independent parts, launch onegeneral-purposeagent per part withmodel: "sonnet"andisolation: "worktree", each told its files, the sketch branch to start from, these hard rules, and to commit and reply with its branch name; then merge their branches into the sketch branch yourself, fix the build, and continue. - Never edit plan, milestone or ledger docs, even to note the sketch. Things worth remembering go in the hand-off summary (step 5).
Steps
- Brief. Ask at most 3 questions, only ones whose answer changes what you build; otherwise pick sensible defaults. Write a brief of 3–6 lines in the chat: what it does, what's left out, how the user will try it.
- Build it — the smallest version that shows the idea working end to end. Reuse existing code; hard-code what a later version would configure.
- Show it. Build, run the touched tests, then launch the app the way the project does (the
runskill if available, else the project's documented run command) and give the user the exact command plus 2–4 things to try and look at. - Iterate. Each piece of feedback → change → build → commit → tell the user what to look at now. Keep going until the user says keep, harden or drop.
- Hand off — when the user decides:
- Keep as is: merge the sketch branch into the branch it came from (
git merge --no-ff sketch/<slug>) after running the stack reviewer the global CLAUDE.md "Agents" section names on the whole diff and fixing its correctness findings. - Harden: leave the branch; write a hand-off summary (what was built, files, decisions made along the way, known gaps and shortcuts) and suggest the next step —
/milestoneor/impl-planwith the summary andgit diff <base>...sketch/<slug>as input. - Drop: switch back to the original branch; offer
git branch -D sketch/<slug>and run it only on a yes.
- Keep as is: merge the sketch branch into the branch it came from (
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.

