feature-workflow
Skill11 starsChanged 12 days agoMITView on GitHub
Standard workflow for developing features. Follow this process for all non-trivial changes - from planning through PR merge. Ensures proper testing, review, and CI verification.
- Deletes or force-pushes
- Commits and pushes
10 other projects use a version of this.
Adopted by
Most recently taken up by
- byjg/skillstheir own version, first seen Nov 1, 2026
- marcowuethrich/green-printtheir 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
- BojanKocijan/bk-charterlinetheir own version, first seen Oct 13, 2026
- cubeplexai/cubeplextheir own version, first seen Sep 28, 2026
Other versions10 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 10
- cubeplexai/cubeplex★ 575same name
- TurboKach/claude-code-setup★ 11same name
- BojanKocijan/bk-charterline★ 4same name
- BojanKocijan/design-forge★ 4same name
- blas1n/claude-skills★ 2same name
- jangjun7091/ai-company★ 1same name
- marcowuethrich/green-print★ 1same name
- byjg/skills★ 0same name
- epinnock/feature-workflow★ 0same name
- otaku1123/claude-code-template★ 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 repository6 other files it gives its agents.
- AGENTS.mdAGENTS.md
- agent-evaluationSkill
- opencode-session-dbSkill
- plugin-testingSkill
- readiness-checkSkill
- opencode-slack-integrationSkill
The file
We split this skill into the rules it gives an agent
12 rules. 0 are common, 12 are distinctive, and 0 are very specific to this project.
0
Common
Most of the 7 edited versions we read keep these.
12
Distinctive
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
- could be automatic: a setting could enforce it
- Distinctive: line 26
Before writing any code, plan the work:
- Distinctive: line 56
Check for existing issue or create one:
- Distinctive: line 90
Update the issue with detailed scope:
- Distinctive: line 121
Always start from up-to-date main:
- Distinctive: line 129
Commit them to current branch
- Distinctive: line 153
Write the code following these principles:
- Distinctive: line 168
Update Todo tool: after each significant piece of work.
- Distinctive: line 193
If tests fail, fix before proceeding: Do not skip failing tests.
- Distinctive: line 305
Read the failure logs
- Distinctive: line 308
Wait for CI again
- Distinctive: line 351
Pushing directly to main: - Always use feature branches
- Distinctive: line 358
Force pushing to main: - Never force push to shared branches
“Common” is measured: kept by at least 6 in 10 of the 7 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 Development Workflow
A structured 11-step process for developing features that ensures quality, traceability, and proper review.
Prerequisites
- Git repository initialized
- GitHub CLI (
gh) authenticated - Access to run tests (unit and E2E)
- Todo tool available for tracking progress
The 11-Step Process
Step 1: Create a Plan
Before writing any code, plan the work:
## Feature: [Name]
### Goal
[One sentence describing what we're building]
### Why
[Problem this solves or value it provides]
### Scope
- [ ] What's included
- [ ] What's NOT included (explicit boundaries)
### Technical Approach
1. [High-level step 1]
2. [High-level step 2]
3. [etc.]
### Risks / Open Questions
- [Any unknowns or concerns]Use the Todo tool to capture each major task from the plan.
Step 2: Create GitHub Issue (if not exists)
Check for existing issue or create one:
# Search for existing issue
gh issue list --repo OWNER/REPO --search "feature keywords"
# Create new issue if needed
gh issue create --repo OWNER/REPO \
--title "feat: [Feature Name]" \
--body "$(cat <<'EOF'
## Summary
[Brief description]
## Motivation
[Why this is needed]
## Proposed Solution
[High-level approach]
## Acceptance Criteria
- [ ] Criterion 1
- [ ] Criterion 2
- [ ] Tests pass
- [ ] Documentation updated
EOF
)"Record the issue number for linking in commits and PR.
Step 3: Define Task Scope in Issue
Update the issue with detailed scope:
gh issue comment ISSUE_NUMBER --body "$(cat <<'EOF'
## Implementation Plan
### Design
[Architecture decisions, data flow, etc.]
### Files to Modify
- `path/to/file1.ts` - [what changes]
- `path/to/file2.ts` - [what changes]
### New Files
- `path/to/new.ts` - [purpose]
### Testing Strategy
- Unit tests for [X]
- E2E tests for [Y]
- Manual verification of [Z]
### Out of Scope
- [Explicitly list what this PR won't do]
EOF
)"Step 4: Fetch Latest Changes
Always start from up-to-date main:
git fetch origin
git status # Check for uncommitted changesIf you have uncommitted changes, either:
- Commit them to current branch
- Stash them:
git stash - Discard them:
git checkout -- .
Step 5: Create Feature Branch
Branch naming convention: feat/issue-number-short-description
# Create and checkout new branch from origin/main
git checkout -b feat/123-add-telegram-replies origin/main
# Or for fixes
git checkout -b fix/456-race-condition origin/mainUpdate Todo tool: Mark "Create branch" as complete.
Step 6: Implement the Feature
Write the code following these principles:
- Small, focused commits - Each commit should be atomic
- Commit message format:
type(scope): short description - Detail 1 - Detail 2 Closes #123 - Types:
feat,fix,refactor,test,docs,chore - Reference issue in commits:
#123orCloses #123
Update Todo tool after each significant piece of work.
Step 7: Test End-to-End (No Mocks)
Critical: Real testing, not mocked data.
# 1. Run type checking
npm run typecheck
# 2. Run unit tests
npm test
# 3. Run E2E tests (REQUIRED for any plugin changes)
OPENCODE_E2E=1 npm run test:e2e
# 4. Manual verification
# - Test the actual feature with real data
# - For Telegram: send real messages
# - For TTS: verify audio plays
# - For reflection: verify judge runsIf tests fail, fix before proceeding. Do not skip failing tests.
Add automated tests for new functionality:
- Unit tests for pure logic
- Integration tests for component interactions
- E2E tests for user-facing flows
Step 8: Create Pull Request
# Ensure branch is pushed
git push -u origin HEAD
# Create PR linking to issue
gh pr create \
--title "feat: [Short description] (#ISSUE)" \
--body "$(cat <<'EOF'
## Summary
[What this PR does]
## Changes
- [Change 1]
- [Change 2]
## Testing
- [ ] Unit tests pass
- [ ] E2E tests pass
- [ ] Manual testing completed
## Screenshots/Logs
[If applicable]
Closes #ISSUE_NUMBER
EOF
)"Step 9: Review PR
Self-review checklist:
# View the full diff
gh pr diff
# Check files changed
gh pr view --json filesReview for:
- No debug code or console.logs left
- No hardcoded secrets or credentials
- Error handling is appropriate
- Code is readable and well-commented
- No unrelated changes included
- Commit history is clean
Clean up if needed:
# Squash fixup commits
git rebase -i origin/main
# Force push after rebase (only on feature branch!)
git push --force-with-leaseStep 10: Update GitHub Issue
Post implementation summary to the issue:
gh issue comment ISSUE_NUMBER --body "$(cat <<'EOF'
## Implementation Complete
### What was done
- [Summary of changes]
### Files changed
- `path/to/file.ts` - [description]
### How to test
1. [Step 1]
2. [Step 2]
### PR
#PR_NUMBER
EOF
)"Step 11: Wait for CI Checks to Pass
# Watch CI status
gh pr checks --watch
# Or check run status
gh run list --limit 5
# View specific run logs if failed
gh run view RUN_ID --log-failedIf CI fails:
- Read the failure logs
- Fix the issue locally
- Push the fix
- Wait for CI again
Only merge when all checks pass.
# Merge when ready (if you have permission)
gh pr merge --squash --delete-branchQuick Reference
| Step | Command | Todo Status |
| 1. Plan | Document in markdown | pending |
| 2. Issue | gh issue create | pending |
| 3. Scope | gh issue comment | pending |
| 4. Fetch | git fetch origin | in_progress |
| 5. Branch | git checkout -b feat/... | in_progress |
| 6. Implement | Write code, commit | in_progress |
| 7. Test | npm test && npm run test:e2e | in_progress |
| 8. PR | gh pr create | in_progress |
| 9. Review | gh pr diff | in_progress |
| 10. Update Issue | gh issue comment | in_progress |
| 11. CI Pass | gh pr checks --watch | completed |
Common Issues
| Issue | Cause | Solution |
| CI fails on push | Tests not run locally | Always run npm test before pushing |
| Merge conflicts | Branch out of date | git rebase origin/main |
| E2E tests timeout | Model too slow | Increase timeout or use faster model |
| PR has unrelated changes | Wrong base branch | Rebase onto correct branch |
| Forgot to link issue | Missing Closes #N | Edit PR body to add it |
Anti-Patterns to Avoid
- Pushing directly to main - Always use feature branches
- Skipping tests - Tests exist for a reason
- Large PRs - Break into smaller, focused changes
- Mocked E2E tests - Real tests catch real bugs
- Ignoring CI failures - Fix before merge, never skip
- No issue tracking - Issues provide context and history
- Vague commit messages - Be specific about what changed
- Force pushing to main - Never force push to shared branches
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.

