feature-workflow
Skill1 starsChanged 17 days agoMITView on GitHub
Takes one GitHub issue of this repository from spec to a merged pull request - worktree, spec, plan, subagents, the every-surface walk, the zero-findings gate, a fresh review, the pull request and the cleanup after the merge. Use when asked to build, implement or ship an issue or a feature here.
11 other projects use a version of this.
Adopted by
Most recently taken up by
- arraydude/agent-skillstheir own version, first seen Nov 16, 2026
- cubeplexai/cubeplextheir own version, first seen Nov 10, 2026
- dzianisv/opencode-pluginstheir own version, first seen Nov 1, 2026
- byjg/skillstheir own version, first seen Nov 1, 2026
- BojanKocijan/design-forgetheir own version, first seen Oct 19, 2026
- otaku1123/claude-code-templatetheir own version, first seen Oct 13, 2026
Other versions11 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 11
- cubeplexai/cubeplex★ 576same name
- dzianisv/opencode-plugins★ 11same name
- TurboKach/claude-code-setup★ 11same name
- BojanKocijan/bk-charterline★ 4same name
- BojanKocijan/design-forge★ 4same name
- arraydude/agent-skills★ 2same name
- blas1n/claude-skills★ 2same name
- jangjun7091/ai-company★ 1same name
- byjg/skills★ 0same name
- epinnock/feature-workflow★ 0same name
1 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.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- kotlin-data-kmp-data-layerSkill
- kotlin-kmp-code-reviewSkill
- kotlin-project-feature-implementationSkill
- kotlin-testing-kmpSkill
- kotlin-ui-compose-multiplatformSkill
The file
We split this skill into the rules it gives an agent
4 rules. 0 are common, 2 are distinctive, and 2 are very specific to this project.
0
Common
Most of the 9 edited versions we read keep these.
2
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
- Distinctive: line 13
Read the issue and the docs AGENTS.md routes it to.
- Very specific: line 38
Test first where the behaviour is clear: a ViewModel test in shared, a Compose UI test in composeApp/src/commonTest, a resource test in …
- Very specific: line 44
Run what changed the way a user would (desktop: ./gradlew :composeApp:run, web: ./gradlew :composeApp:wasmJsBrowserDevelopmentRun, Android …
- Distinctive: line 60
Remove the worktree and the branch (git worktree remove, git branch -D, the remote branch).
“Common” is measured: kept by at least 6 in 10 of the 9 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
Feature workflow
The repository's way to land a change. Every step names its evidence; a step without evidence is not done.
1. Start from the issue
- Read the issue and the docs
AGENTS.mdroutes it to. Unclear scope: ask now with the question tool, with a recommendation, never later in prose. - One worktree per issue, never
main:git worktree add .agents/worktrees/<n>-<slug> -b feat/<n>-<slug> origin/main.
2. Spec first
- A change a user sees gets its spec in
docs/features/<name>.md: user stories, flows, edge cases, the reverse states (how to undo or leave it), what is out of scope. - An architectural choice gets an ADR in
docs/architecture/decisions.md, in the same pull request.
3. Plan, then delegate
- Split the work into tasks that each stand alone: the shared model and repository, the server endpoint, the UI, the tests.
- A task with a clear spec goes to a subagent with the whole brief (it sees nothing else): the
issue, the worktree path, the files it owns, the rules from
AGENTS.mdit must keep, what to return. Choose its model and effort asdocs/agentic/README.md(Subagents) says. - The lead keeps the integration, the questions and the pull request.
4. Build with the project's skills
kotlin-project-feature-implementationfor placing the code,kotlin-data-kmp-data-layerfor repositories and sync,kotlin-ui-compose-multiplatformfor the UI,kotlin-testing-kmpfor the tests.- Test first where the behaviour is clear: a ViewModel test in
shared, a Compose UI test incomposeApp/src/commonTest, a resource test inserver.
5. Prove it
- Walk Hit every surface in
AGENTS.mdand write down which entries applied. - Run what changed the way a user would (desktop:
./gradlew :composeApp:run, web:./gradlew :composeApp:wasmJsBrowserDevelopmentRun, Android or iOS on a simulator) and keep a screenshot for the pull request. ./gradlew prePRcheckonce, clean. A finding gets fixed, never baselined.
6. Review and pull request
- A fresh reviewer that did not write the code reads the diff (
kotlin-kmp-code-review), with the code, not the author's conclusions. Fix what holds up. - Open the pull request with the template:
Closes #<n>, the surfaces, the evidence, the gate. - Watch the checks; fix a failure in the same branch. Merge when every check is green and every
thread is resolved, never with
--admin.
7. Clean up
- Fast-forward the main checkout to
origin/main. - Remove the worktree and the branch (
git worktree remove,git branch -D, the remote branch). - File what stayed open as a new issue, so nothing lives only in a chat.
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.

