dev-workflow
Skill4 starsChanged 7 months agoMITView on GitHub
Master orchestrator for the full autonomous development lifecycle. Use this skill whenever the user mentions starting a new feature, has an idea to implement, wants to plan development, says "quero desenvolver", "tenho uma ideia", "nova feature", "começar uma issue", "próxima issue", or asks what the next step in development is. This skill coordinates all phases and delegates to existing skills (PRD, Issues, Code Review) where available, filling in QA Planning, Development, Testing, and PR management. Trigger even if the user only mentions one phase — this skill decides what comes next and runs autonomously via Claude Code.
32 other projects use a version of this.
Adopted by
Most recently taken up by
- AsiaOstrich/EngramGraphtheir own version, first seen Nov 8, 2026
- pandazki/pneuma-skillstheir own version, first seen Nov 6, 2026
- bmsuisse/skillstheir own version, first seen Nov 6, 2026
- openflocon/Flocontheir own version, first seen Nov 1, 2026
- Youngju-Heo/claude-skilltheir own version, first seen Nov 1, 2026
- dzianisv/skillstheir own version, first seen Nov 1, 2026
Used with
Projects with it also use retrospective, refactor and 4 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.
MicrosoftSkillretrospective
Review how work went and improve the next round. USE WHEN asked for a retro, delays, or process lessons. DO NOT USE for code review or transcript replay.
Also in 3 projects that use this
GitHubSkillrefactor
Surgical code refactoring to improve maintainability without changing behavior. Covers extracting functions, renaming variables, breaking down god functions, improving type safety, eliminating code smells, and applying design patterns. Less drastic than repo-rebuilder; use for gradual improvements.
Also in 4 projects that use this
HashiCorpSkillchangelog
Add a `.changelog/<PR_NUMBER>.txt` entry for a GitHub Pull Request, commit, and push.
Also in 3 projects that use this
VercelSkilldeploy
Deploy applications to Vercel. Use when the user says "deploy", "deploy to Vercel", "push to production", "deploy my app", or "go live".
Also in 3 projects that use this
GoogleSkilltdd
Test-driven development. Use when the user wants to build features or fix bugs test-first, mentions "red-green-refactor", or wants integration tests.
Also in 4 projects that use this
SentrySkillcommit
Create commit messages following Sentry conventions. Use when committing code changes, writing commit messages, or formatting git history. Follows conventional commits with Sentry-specific issue references.
Also in 3 projects that use this
Other versions32 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 32
- ag2ai/faststream★ 5.4ksame name
- xstongxue/best-skills★ 2.9ksame name
- AltimateAI/vscode-dbt-power-user★ 586same name
- openflocon/Flocon★ 379same name
- SmartThingsCommunity/SmartThingsEdgeDrivers★ 349same name
- pandazki/pneuma-skills★ 161same name
- AsiaOstrich/universal-dev-standards★ 84same name
- clarity-digital-development/tworkflow★ 47same name
- addxai/enterprise-harness-engineering★ 44same name
- muyen/vibe-to-prod★ 34same name
22 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 repository21 other files it gives its agents.
- analyticsSkill
- brainstormSkill
- brandingSkill
- code-reviewSkill
- devops-deploySkill
- founderSkill
- grill-meSkill
- idea-validationSkill
- improve-codebase-architectureSkill
- landing-pageSkill
- legal-complianceSkill
- marketingSkill
- prd-to-issuesSkill
- senior-backendSkill
- senior-frontendSkill
- senior-qaSkill
- seoSkill
- software-architectSkill
- video-creatorSkill
- video-editorSkill
- write-a-prdSkill
The file
Show its rules and text
We split this skill into the rules it gives an agent
19 rules. 0 are common, 17 are distinctive, and 2 are very specific to this project.
0
Common
Most of the 23 edited versions we read keep these.
17
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 65
If starting fresh, ask: "Nova feature ou continuando uma issue existente?"
- Distinctive: line 74
Wait for the PRD to be submitted as a GitHub issue before advancing.
- Distinctive: line 84
If the PRD contains ambiguities, run a structured clarification pass: identify up to 5 critical unknowns, resolve each through targeted Q&A …
- Distinctive: line 92
Ask all questions in a single message.
- Distinctive: line 93
Update the PRD issue with the resolved answers and mark the clarification pass complete as a comment.
- Distinctive: line 96
Skip this phase: when the PRD is unambiguous or the user explicitly says to proceed.
- Distinctive: line 98
Exit condition: All critical unknowns resolved and documented on the PRD issue.
- Distinctive: line 109
Wait for all issues to be created and confirmed by the user before advancing.
- Distinctive: line 111
Exit condition: All issues created on GitHub, order confirmed by user.
- Distinctive: line 159
Exit condition: Test plan posted as issue comment.
- Very specific: line 170
Use senior-frontend for UI components, pages, and client-side logic.
- Very specific: line 171
Use senior-backend for API routes, database schemas, and server-side logic.
- Distinctive: line 191
Always use npm to run tests, regardless of what package manager the project uses:
- Distinctive: line 205
Run smoke tests on every feature listed in "Regression Scope" from the QA plan.
- Distinctive: line 216
Exit condition: All green, coverage >= 80%, zero regressions.
- Distinctive: line 226
Wait for verdict before advancing.
- Distinctive: line 258
Check off the corresponding item on the parent PRD issue
- Distinctive: line 262
All issues done -> close parent PRD issue with summary -> workflow complete
- Distinctive: line 293
Never skip a gate: If the user requests skipping, state the risk clearly and require explicit confirmation before proceeding.
“Common” is measured: kept by at least 6 in 10 of the 23 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
Dev Workflow Orchestrator
Autonomous end-to-end lifecycle manager for Claude Code. Coordinates all phases from idea to merged PR. Delegates to your existing skills where they exist; owns the phases in between.
Skill Map
| Phase | Owner |
| 1 — PRD | Invoke skill: write-a-prd |
| 2 — Issues from PRD | Invoke skill: prd-to-issues |
| 3 — Architecture | Invoke skill: software-architect |
| 4 — QA Planning | This skill |
| 5 — Development | Invoke skills: senior-frontend, senior-backend |
| 6 — Automated Testing | This skill (npm test + Playwright) |
| 7 — Code Review | Invoke skill: code-review:code-review |
| 8 — PR + Close + Advance | This skill |
Workflow
IDEA
|-- [Phase 1] invoke skill: write-a-prd
|-- [Phase 2] invoke skill: prd-to-issues -> ordered issues
|-- For each issue:
|-- [Phase 3] invoke skill: software-architect
|-- [Phase 4] QA Planning (before any code)
|-- [Phase 5] invoke skills: senior-frontend + senior-backend
|-- [Phase 6] npm test + Playwright + coverage check
|-- [Phase 7] invoke skill: code-review:code-review
|-- [Phase 8] PR created -> issue closed -> next issueSession Start Protocol
At the start of every session, output this state block and ask the user to confirm:
DEV WORKFLOW STATE
Feature PRD : #[number] - [title]
Current issue : #[number] - [title]
Current phase : [1-8] - [phase name]
Remaining : #X, #Y, #ZIf starting fresh, ask: "Nova feature ou continuando uma issue existente?"
Phase 1 — PRD
Owner: your existing skill.
Invoke skill write-a-prd to create the PRD through user interview and codebase exploration.
Wait for the PRD to be submitted as a GitHub issue before advancing.
Exit condition: PRD issue created on GitHub.
Phase 1.5 — Clarify
Trigger: PRD exists but contains ambiguities before issue breakdown begins.
If the PRD contains ambiguities, run a structured clarification pass: identify up to 5 critical unknowns, resolve each through targeted Q&A before proceeding to Phase 2.
- Scan the PRD for gaps: missing edge-case behaviour, undefined user roles, vague success criteria, unstated constraints, or conflicting requirements.
- List each unknown as a numbered question (max 5). Prioritise questions that affect scope, security, or core UX — skip cosmetic or low-stakes ambiguities.
- Ask all questions in a single message. Wait for answers before continuing.
- Update the PRD issue with the resolved answers and mark the clarification pass complete as a comment.
Skip this phase when the PRD is unambiguous or the user explicitly says to proceed.
Exit condition: All critical unknowns resolved and documented on the PRD issue.
Phase 2 — Issues from PRD
Owner: your existing skill.
Invoke skill prd-to-issues to break the PRD into independently-grabbable GitHub issues
using tracer-bullet vertical slices. The skill handles ordering, dependencies, and scoping.
Wait for all issues to be created and confirmed by the user before advancing.
Exit condition: All issues created on GitHub, order confirmed by user.
Phase 3 — Architecture
Owner: software-architect skill.
Invoke skill software-architect to design the system structure for this issue/feature.
For the first issue in a new feature, this produces the full architecture. For subsequent
issues, review and update the existing architecture if the new issue changes boundaries.
Skip this phase for small issues that don't affect system structure (bug fixes, copy changes, minor UI tweaks).
Exit condition: Architecture document exists and is appropriate for the issue scope.
Phase 4 — QA Planning
Trigger: Architecture exists (or skipped) for this issue. Rule: This runs BEFORE any code is written.
Generate and post as a GitHub comment on the issue:
## Test Plan - #[issue-number]
### Unit Tests
- [ ] [unit] should [behavior] when [condition]
### Integration Tests
- [ ] [flow] should [result] given [state]
### Edge Cases
- [ ] Null / empty inputs
- [ ] Boundary values
- [ ] Concurrent operations (if applicable)
- [ ] Network / auth failures (if applicable)
### E2E Scenarios (Playwright)
- [ ] Happy path: [user does X -> sees Y]
- [ ] Error path: [user does X -> sees error Z]
### Regression Scope
- Features this change touches: [list files/components]Exit condition: Test plan posted as issue comment. Do not write code before this exists.
Phase 5 — Development
Trigger: QA plan exists on the issue.
Invoke skills senior-frontend and/or senior-backend as appropriate for the implementation.
Use senior-frontend for UI components, pages, and client-side logic.
Use senior-backend for API routes, database schemas, and server-side logic.
Claude Code implements autonomously following these rules:
- Never commit directly to
main - Branch name:
feat/#[issue-number]-short-description - Commit format:
feat: [description] (#[issue-number]) - No TODOs left in code
- Implement against ACs — keep the test plan visible while coding
Exit condition: Implementation complete, all files saved, ready for testing.
Phase 6 — Automated Testing
Trigger: Implementation complete.
5a — Unit & Integration
Always use npm to run tests, regardless of what package manager the project uses:
npm test -- --coverageGate: coverage >= 80%. If below, return to Phase 5 with coverage report.
5b — E2E via Playwright Agents (three-agent flow)
- Planner — converts QA plan E2E scenarios into navigation flows
- Generator — writes Playwright tests, validates selectors against running app
- Healer — if a test fails, replays, diagnoses, and patches automatically
5c — Regression
Run smoke tests on every feature listed in "Regression Scope" from the QA plan.
Test Report (post as issue comment)
## Test Results - #[issue-number]
- Unit/Integration : X passing | X failing
- Coverage : X% (gate: 80%)
- E2E : X passing | X failing
- Regressions : None | [list]Exit condition: All green, coverage >= 80%, zero regressions. On failure, return to Phase 5 with the report attached.
Phase 7 — Code Review
Owner: your existing Code Review skill.
Invoke skill code-review:code-review directly. Pass it the branch/diff context.
Wait for verdict before advancing.
- Approved -> advance to Phase 8
- Changes requested -> return to Phase 5 with review comments
Exit condition: Code Review skill returns Approved.
Phase 8 — PR + Close + Advance
Trigger: Code review approved.
7a — Create PR
## What
[One paragraph describing the change]
## Why
Closes #[issue-number]
Part of #[parent-PRD-number]
## Checklist
- [x] Unit/Integration tests passing
- [x] E2E tests passing
- [x] Coverage >= 80%
- [x] No regressions detected
- [x] Code review approved7b — After merge
- Close issue with comment:
Resolved via #[PR-number] - Check off the corresponding item on the parent PRD issue
7c — Advance
- Issues remaining -> update state block -> return to Phase 3 (or Phase 4 if architecture unchanged) for next issue
- All issues done -> close parent PRD issue with summary -> workflow complete
Exit condition: PR merged, issue closed, next phase identified.
Reference Templates
This skill has detailed reference templates. Consult references/spec-templates.md
for structured templates for specs, plans, tasks, and checklists.
| Template | Purpose | When to use |
| Feature Specification | Full spec with scenarios, FRs, entities, success criteria | Phase 1 PRD creation |
| Implementation Plan | Phased plan with research, gates, MVP cut line | Phase 2 issue breakdown |
| Task Breakdown | T-NNN task list with parallel markers and user story links | Phase 2 / Phase 4 prep |
| Requirements Checklist | CHK-NNN quality gates for spec approval | Before Phase 2 begins |
| Constitution (optional) | Project-level architectural principles | Phase 0 for greenfield projects |
Quality Gates
| Gate | Requirement | On failure |
| QA Plan exists | Before Phase 5 starts | Block — do not proceed |
| Unit coverage | >= 80% | Return to Phase 5 |
| E2E tests | All passing | Return to Phase 5 |
| Regressions | Zero detected | Return to Phase 5 |
| Code Review | Approved | Return to Phase 5 |
Never skip a gate. If the user requests skipping, state the risk clearly and require explicit confirmation before proceeding.
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.

