development-workflow（开发工作流程）

Development Workflow

Structured approach to software development ensuring requirements are clearly defined, designs are meticulously planned, and implementations are thoroughly documented with proper contribution practices.

• Leverage native parallel subagent dispatch and 200k+ context windows where available.

Activation Conditions

Use symptom -> action triggers: when one matches, apply this skill and verify with the protocol below.

Project Planning & Requirements:

• Starting new features or phases of work
• Defining requirements using structured notation
• Creating technical designs and architecture documents
• Managing implementation plans from concept to completion
• Ensuring thorough documentation before coding
• Understanding project structure and existing patterns

Repository Contributions:

• Creating issues, commits, or pull requests in a repository
• Needing to follow repository guidelines before making contributions
• Creating PRs, pushing code, or following contribution workflows
• Following project-specific contribution guidelines

Quality Assurance:

• Ensuring two-stage review (spec compliance first, then code quality) standards are met
• Validating tests and requirements before implementation
• Managing project decisions and trade-offs documentation
• Tracking progress and blockers

Part 1: Spec-Driven Development

Core Artifacts

Maintain these artifacts throughout the project lifecycle:

EARS Notation for Requirements

EARS = Easy Approach to Requirements Syntax

Basic EARS Patterns

1. Universal Requirements Apply to all entities without condition

• "The system shall validate all user input."
• "Each user shall have a unique email address."

2. State-Driven Requirements Apply only in specific system states

• "When the user is authenticated, the system shall display the dashboard."
• "If the payment fails, the system shall retry up to 3 times."

3. Event-Driven Requirements Triggered by specific events

• "When the user clicks 'Submit', the system shall validate the form."
• "Upon receiving a new message, the chat application shall update the conversation view."

4. Optional-Feature Requirements Describing optional or conditional features

• "The system may provide offline access if supported."
• "The user may choose to receive email notifications."

5. Unwanted Behavior Specifying what should not happen

• "The system shall not store passwords in plain text."
• "The application shall not allow simultaneous sessions from different locations unless configured."

6. AI-Agent Requirements Specifying agent-visible constraints, handoff evidence, and autonomous execution limits

• "The agent shall run the documented verification command before claiming completion."
• "The agent shall not modify files outside the task scope without explicit approval."

Complete EARS Example

# User Authentication Requirements

## Universal Requirements
- U-001: The system **shall** require users to provide an email address and password for login.
- U-002: The system **shall** validate email addresses using RFC 5322 format.

## Event-Driven Requirements
- E-001: **When** the user clicks "Forgot Password", the system **shall** send a password reset link to the registered email.
- E-002: **Upon** successful authentication, the system **shall** generate a session token valid for 24 hours.

## State-Driven Requirements
- S-001: **If** the user has enabled two-factor authentication, the system **shall** prompt for the verification code.
- S-002: **When** the account is locked due to too many failed attempts, the system **shall** unlock it after 30 minutes.

## Unwanted Behavior
- N-001: The system **shall not** reveal whether an email address is registered during password reset.
- N-002: The system **shall not** allow the same account to be used from more than 3 IP addresses simultaneously.

## Optional Features
- O-001: The system **may** support social login providers (Google, Facebook, GitHub).

## AI-Agent Requirements
- A-001: The agent **shall** report changed files, commands run, and unresolved risks before handoff.

EARS Grammar and Parser Example

requirement = id ":" [trigger ","] subject "shall" action "." ;
trigger = ("When" | "If" | "While" | "Upon") condition ;
id = ("U" | "E" | "S" | "N" | "O" | "A") "-" digit digit digit ;

const earsRequirement =
  /^(?<id>[UESNOA]-\d{3}):\s(?:(?<keyword>When|If|While|Upon)\s(?<condition>.+?),\s)?(?<subject>The system|The user|The agent)\sshall\s(?<action>.+)\.$/;

Overview

High-level description of what this feature does and why it's needed.

Architecture

System Diagram

User → Frontend → API Gateway → Service → Database
                              ↓
                         Cache Layer

Component Structure

src/
├── components/
│   └── FeatureName/
│       ├── FeatureComponent.tsx
│       ├── SubComponent.tsx
│       └── styles.css
├── services/
│   └── featureService.ts
├── api/
│   └── featureApi.ts
└── types/
    └── feature.types.ts

Data Models

entities/Feature.ts

interface FeatureEntity {
  id: string;
  name: string;
  status: 'active' | 'inactive';
  createdAt: Date;
  // ... other fields
}

API Contracts

POST /api/features

interface CreateFeatureRequest {
  name: string;
  /* ... other fields */
}

interface CreateFeatureResponse {
  id: string;
  status: 'created';
}

Error Handling

Sequence Diagram

User → Frontend: Click "Create Feature"
Frontend → API: POST /api/features
API → Validator: Validate data
Validator → API: Valid / Invalid
API → Database: Insert feature
Database → API: Created
API → Frontend: Return feature ID
Frontend → User: Show success message

Security Considerations

• Authentication required for all mutations
• Input validation on all endpoints
• Rate limiting on create operations
• Audit logging for all changes

Performance Considerations

• Caching strategy for read operations
• Database indexing requirements
• CDN for static assets

Agentic Considerations

• Files or directories agents may modify
• Commands agents must run before completion
• Decisions that require human approval
• Handoff evidence required for review

Implementation Phases

1. Phase 1: Core CRUD operations
2. Phase 2: Validation and error handling
3. Phase 3: Caching layer
4. Phase 4: Testing and documentation

### Implementation Task Tracking

Tasks: [Feature Name] Implementation

Phase 1: Foundation

• [ ] Confirm scope, dependencies, and agent boundaries
• [ ] Create data model and migration scaffold

Phase 2: Core Functionality

• [ ] Implement core API or service path
• [ ] Add validation, errors, and persistence

Phase 3: Frontend Integration

• [ ] Build main component
• [ ] Cover loading, empty, error, and success states

Phase 4: Verification

• [ ] Add unit/integration/component tests for changed paths
• [ ] Run required verification commands and record evidence

Phase 5: Release Notes

• [ ] Document behavior, migration, and user-facing changes

Full scaffold reference: examples/feature-spec-example.md or scripts/create-spec-scaffold.ps1.

---

## Part 2: Repository Contribution Guidelines

## Anti-Patterns

- Starting work before the plan or gate is clear: Execution drifts when success criteria are implied instead of explicit.
- Treating verification as optional cleanup: The last mile is where regressions and missing updates are usually hiding.
- Mixing planning, implementation, and release work in one jump: You lose the causal chain that explains why a change is safe.

## Verification Protocol

Before claiming "skill applied successfully":

1. Pass/fail: The Development Workflow starts from explicit success criteria, constraints, and stop conditions.
2. Pass/fail: Required evidence is collected before any completion, approval, or readiness claim.
3. Pass/fail: The next action follows the documented gate order without skipping review or verification steps.
4. Pressure-test scenario: Apply the workflow under time pressure with one failing check and one tempting shortcut.
5. Success metric: Zero rationalizations; blocked, failed, or unverified work is reported as such.

### Pre-Contribution Checklist

Before creating any PR or making changes:

Pre-Submission Checklist

Repository Understanding

• [ ] Read README.md for project overview
• [ ] Read CONTRIBUTING.md for contribution rules
• [ ] Reviewed issue and PR templates
• [ ] Checked for existing related issues or PRs

Environment Setup

• [ ] forked repository (if required)
• [ ] cloned repository locally
• [ ] setup development environment
• [ ] installed all dependencies

Development Standards

• [ ] Understand coding style conventions
• [ ] Familiar with branching strategy
• [ ] Aware of commit message conventions
• [ ] Know testing requirements

### Issue Creation Workflow

#### When to Create Issues

- New feature requests
- Bug reports
- Documentation gaps
- Security vulnerabilities
- Performance concerns

#### Issue Template

Issue: [Brief Title]

Type: [Feature Request / Bug Report / Question / Documentation]

Priority: [Low / Medium / High / Critical]

Description

[Detailed description of the issue or feature request]

Reproduction Steps (for bugs)

1. [First step]
2. [Second step]
3. [Third step]

Expected Behavior

[What should happen]

Actual Behavior

[What actually happened]

Environment

• OS: [e.g., Windows 10, macOS 12.5]
• Browser: [e.g., Chrome 120, Firefox 115]
• Version: [If applicable]

Additional Context

[Any other information that might be helpful]

• Screenshots
• Error messages
• Logs

### Branching Strategy

Branch Naming Convention

Feature Branches

feature/[JIRA-TICKET]/[short-description] Examples:

• feature/PROJ-123/user-authentication
• feature/PROJ-456/dark-mode-support

Bugfix Branches

bugfix/[JIRA-TICKET]/[issue-description] Examples:

• bugfix/PROJ-789/login-redirect-loop
• bugfix/PROJ-321/memory-leak-report

Hotfix Branches (production issues)

hotfix/[JIRA-TICKET]/[critical-description] Examples:

• hotfix/PROJ-999/security-vulnerability
• hotfix/PROJ-888/database-down

Release Branches

release/v[major].[minor] Examples:

• release/v1.2.0
• release/v2.0.0

Branch Protection Rules

• Protect main and develop branches
• Require pull request reviews (at least 1 approval)
• Require status checks to pass (CI builds, tests)
• Require branches to be up-to-date before merging

### Commit Message Standards (Conventional Commits)

Conventional Commit Format

[optional scope]:

[optional body]

[optional footer(s)]

Commit Types

Breaking Changes

feat(api)!: remove deprecated v1 endpoint

BREAKING CHANGE: v1 endpoints are no longer supported. Use v2.

Good Commit Examples

feat(auth): add refresh token mechanism

- Implement JWT refresh tokens
- Add storage for access tokens
- Update token validation logic

fix(ui): resolve mobile navigation issue

- Mobile menu was not closing after clicking links
- Added event listener to handle link clicks
- Tested on iOS Safari and Chrome Mobile

docs(readme): update installation instructions

- Clarified Node.js version requirement
- Added troubleshooting section

Bad Commit Examples (Don't use)

update 2 fixed bug wip changes final work in progress

Commit Workflow

# Stage changes
git add .

# Interactive staging (optional)
git add -i

# Commit with good message
git commit -m "feat(auth): implement OAuth2 login"

# Or use multi-line for body and footer
git commit -m "fix(db): optimize query performance

Added composite index on user_email and created_at
Reduced query time from 500ms to 50ms

Closes issue #123"

Part 3: Pull Request Process

Creating Pull Requests

PR Checklist

## Pull Request Checklist

### Before Submitting
- [ ] Branch is up-to-date with target branch
- [ ] Code compiles without errors
- [ ] All tests pass locally
- [ ] Linters pass without errors
- [ ] Documentation updated (README, API docs, code comments)
- [ ] Self-review completed

### PR Description
- [ ] Clear title following conventional commits
- [ ] Description explains "why" not just "what"
- [ ] Screenshots for UI changes included
- [ ] Breaking changes documented
- [ ] Related issues referenced (e.g., "Closes #123")
- [ ] Testing instructions provided

### Code Quality
- [ ] Follows project coding standards
- [ ] No commented-out code
- [ ] No console.log statements
- [ ] meaningful variable and function names
- [ ] No sensitive data exposed

### Testing
- [ ] New features tested
- [ ] Bug fixes verified
- [ ] Edge cases covered
- [ ] Manual testing completed as described

PR Template

## Description
[Provide a brief description of the changes in this PR]

## Type of Change
- [ ] Bug fix
- [ ] New feature
- [ ] Breaking change
- [ ] Documentation
- [ ] Refactoring
- [ ] Performance improvement

## Issue
[Closes #xxx] or [Related to #xxx]

## Changes Made
### Main Changes
- [Describe the main changes implemented]

### Technical Details
[Any technical details reviewers should know]

### Breaking Changes
[Any breaking changes and migration steps]

## Testing
### Manual Testing Steps
1. [Step 1]
2. [Step 2]

### Automated Testing
- [ ] Unit tests added and passing
- [ ] Integration tests added and passing
- [ ] E2E tests added and passing

## Screenshots
[Include before/after screenshots for UI changes]

## Checklist
- [ ] I have read the contributing guidelines
- [ ] My code follows the project style guidelines
- [ ] I have performed a self-review
- [ ] I have commented my code appropriately
- [ ] Changes have no linting errors
- [ ] All tests are passing
- [ ] Changes do not break existing functionality

two-stage review (spec compliance first, then code quality) Guidelines

For Reviewers

## two-stage review (spec compliance first, then code quality) Principles

### Be Constructive
- Focus on improving code, not criticizing it
- Provide specific examples and suggestions
- Ask questions rather than making demands
- Acknowledge good code and smart solutions

### Review Focus Areas

#### Correctness
- [ ] Logic is correct and handles edge cases
- [ ] Error handling is adequate
- [ ] No obvious bugs or race conditions
- [ ] API contracts are maintained

#### Code Quality
- [ ] Code is readable and maintainable
- [ ] Naming follows conventions
- [ ] Complexity is reasonable
- [ ] No code duplication

#### Architecture
- [ ] Follows established patterns
- [ ] Good separation of concerns
- [ ] Proper abstractions used
- [ ] No tight coupling

#### Security
- [ ] No security vulnerabilities
- [ ] Input validation present
- [ ] Authentication/authorization correct
- [ ] No sensitive data exposure

### Review Priority Comments
🔴 **Blocker**: Must be fixed before merge
🟡 **Concern**: Discuss before merge
🟢 **Suggestion**: Nice to have, not blocking

Part 4: Documentation Standards

Documentation Templates

Action Documentation

### [TYPE] - [ACTION] - [TIMESTAMP]

**Objective**: [Goal being accomplished]

**Context**: [Current state, requirements, reference to prior steps]

**Decision**: [Approach chosen and rationale]

**Execution**: [Steps taken with parameters and commands]

**Output**: [Complete results, logs, metrics]

**Validation**: [Success verification and results]

**Next**: [Continuation plan to next action]

Decision Record

### Decision - [TIMESTAMP]

**Decision**: [What was decided]

**Context**: [Situation and driving data]

**Options**:
- Option 1: [Description] - [Pros/Cons]
- Option 2: [Description] - [Pros/Cons]
- Option 3: [Description] - [Pros/Cons]

**Rationale**: [Why selected option is superior]

**Impact**: [Anticipated consequences]

**Review**: [Reassessment conditions]

Summary Formats

Streamlined Action Log (for changelogs): [TYPE][TIMESTAMP] Goal: [X] → Action: [Y] → Result: [Z] → Next: [W]

Quick Summary (for updates): What: [Brief description] Why: [Context/rationale] How: [Approach taken] Status: [Current state] Next: [Upcoming step]

Dependency Management

• Pin the runtime, package manager, and critical libraries in the design doc before implementation starts.
• Record upgrade constraints, required codemods, and any package that needs special review because it changes build, auth, or persistence behavior.
• Prefer a small scheduled update cadence over infrequent large jumps so the test surface stays understandable.

Vulnerability Scanning

• Run dependency scans as part of the normal quality gate, not only before release.
• Triage findings by exploitability and reachability instead of blindly applying every patch immediately.
• Track accepted risks with owners and revisit dates when a fix cannot ship in the current cycle.

Accessibility Testing

• Add keyboard, focus-order, semantic HTML, and screen-reader expectations to the feature definition instead of treating accessibility as a polishing pass.
• Pair automated tooling such as axe or Playwright accessibility checks with manual review of the critical flows.
• Keep accessibility bugs inside the same Definition of Done as functional bugs for the feature.

Internationalization (i18n) Considerations

• Identify translatable strings, locale-aware formatting, and right-to-left layout risks during planning.
• Keep UI copy, date or number formatting, and fallback language behavior outside hard-coded components.
• Verify overflow, truncation, and validation messages in at least one non-default locale before release.

Part 5: Quality Gates

Cross-Skill Gates

• Use systematic-debugging when a build, test, or rollout gate fails and the root cause is not yet proven.
• Use code-quality when the feature works but the gate is blocked by maintainability, readability, duplication, or review-quality concerns.

Definition of Done

## Definition of Done (DoD)

### Development
- [ ] Code written and committed
- [ ] Code follows project style guide
- [ ] Code compiles without errors
- [ ] No console errors or warnings
- [ ] Self-review completed

### Testing
- [ ] Unit tests written and passing
- [ ] Integration tests written and passing
- [ ] Manual testing completed
- [ ] Edge cases identified and handled
- [ ] Test coverage meets project threshold (>80%)

### Documentation
- [ ] README updated if API changed
- [ ] API documentation updated
- [ ] Code comments added/updated
- [ ] User documentation updated if needed

### two-stage review (spec compliance first, then code quality)
- [ ] Pull request created
- [ ] At least one approval received
- [ ] All review comments addressed
- [ ] CI/CD pipeline passes
- [ ] Merged to appropriate branch

### Deployment
- [ ] Deployment tested in staging
- [ ] Smoke tests passed
- [ ] Performance validated
- [ ] Monitoring/alerting configured
- [ ] Rollback procedure documented

Quality Checklist

## Code Quality Checklist

### Functionality
- [ ] All acceptance criteria met
- [ ] Requirements from design document fulfilled
- [ ] Edge cases handled
- [ ] Error states handled gracefully

### Usability
- [ ] User-friendly error messages
- [ ] Clear feedback for actions
- [ ] Accessible (keyboard, screen reader)
- [ ] Responsive design verified

### Performance
- [ ] Meet performance requirements
- [ ] No memory leaks
- [ ] Efficient algorithms used
- [ ] Appropriate caching

### Security
- [ ] Input validation
- [ ] Output encoding (prevent XSS)
- [ ] Authentication/authorization implemented
- [ ] No sensitive data in logs
- [ ] Dependencies audited and up-to-date

### Maintainability
- [ ] Code is readable
- [ ] Good variable/function names
- [ ] Appropriate abstractions
- [ ] No code duplication
- [ ] Adequate comments

### Testing
- [ ] Unit tests for critical paths
- [ ] Integration tests for API
- [ ] E2E tests for user flows
- [ ] Tests are reliable and not flaky

Part 6: Project Lifecycle Management

Phase-based Development

## Development Phases

### Phase 1: Requirements & Planning
- Gather requirements from stakeholders
- Document in EARS notation
- Create technical design
- Identify risks and dependencies
- Estimate effort and timeline

### Phase 2: Development
- Set up feature branches
- Implement core functionality
- Write tests alongside code
- two-stage review (spec compliance first, then code quality) and iteration

### Phase 3: Testing & QA
- Run automated tests
- Perform manual testing
- Conduct QA review
- Fix identified issues
- Performance testing

### Phase 4: Deployment
- Deploy to staging
- Execute smoke tests
- Get stakeholder approval
- Deploy to production
- Monitor for issues

### Phase 5: Post-Release
- Monitor metrics and logs
- Collect user feedback
- Address critical bugs
- Document lessons learned
- Update process improvements

Development Workflow Best Practices

Before Coding

• Requirements clearly defined and reviewed
• Technical design documented
• Dependencies identified
• Acceptance criteria established
• Testing strategy planned

During Development

• Small, frequent commits
• Write tests alongside code
• Follow coding standards
• Continuous self-review
• Request early feedback

Before Release

• All tests passing
• two-stage review (spec compliance first, then code quality) completed
• Documentation updated
• Security review (if needed)
• Release notes prepared

After Release

• Monitor production carefully
• Be ready to rollback
• Document issues and fixes
• Update metrics and dashboards
• Conduct retrospective

References & Resources

Documentation

• EARS Notation Reference — Complete EARS requirement patterns with examples and traceability matrix
• Design Doc Guide — Technical design document writing guide with ADR template

Scripts

• Create Spec Scaffold — PowerShell script to generate requirements, design, and tasks documents

Examples

• Feature Spec Example — Complete spec-driven development example for User Authentication

Cross-Client Portability

This skill is written to stay usable across GitHub Copilot, Claude Code, Codex, and Gemini CLI.

• GitHub Copilot: keep the folder in a Copilot-visible skill or plugin path, or wrap the workflow as project instructions if the host does not support portable skill folders directly.
• Claude Code: keep the folder in a local skills directory or a compatible plugin or marketplace source.
• Codex: install or sync the folder into $CODEX_HOME/skills/<skill-name> and restart Codex after major changes.
• Gemini CLI: this repository generates a project command named /skills:development-workflow from this skill. Rebuild commands with python scripts/export-gemini-skill.py development-workflow and then run /commands reload inside Gemini CLI.

MCP Availability And Fallback

Preferred MCP Server: None required

• Fallback prompt: "Use the Development Workflow skill without MCP. Rely on the local SKILL.md, bundled references or scripts, and manual verification. Show the exact commands, evidence, and final checks you used before concluding."
• If the current host does not expose a matching server, use the bundled references, scripts, native toolchain, and manual workflow already described in this skill.
• Treat direct local verification, rendered output, logs, tests, or screenshots as the fallback evidence path before completion.

Related Skills

• code-quality: Use it when the workflow also needs two-stage review (spec compliance first, then code quality), maintainability, and refactoring guidance.
• systematic-debugging: Use it when the workflow also needs root-cause debugging before proposing fixes.
• test-driven-development: Use it when the workflow also needs test-first implementation and regression safety.
• verification-before-completion: Use it when the workflow also needs final evidence checks before claiming completion.