| name | feature-dev |
| description | Feature development workflow with exploration, architecture, implementation, and review phases. Use for implementing new features or significant changes. Trigger when user says "implement this feature", "add this functionality", "build this", or describes a feature they want built. |
| metadata | {"argument-hint":"<feature-description>","type":"workflow","agents":[{"name":"code-reviewer","file":"agents/code-reviewer.md","shared":false}]} |
| allowed-tools | Read Write Edit Glob Grep Bash |
Feature Development Workflow
Execute a structured 7-phase feature development workflow. This workflow guides you through understanding, exploring, designing, implementing, and reviewing a feature.
CRITICAL: Complete ALL 7 phases. The workflow is not complete until Phase 7: Summary is finished. After completing each phase, immediately proceed to the next phase without waiting for user prompts.
Phase Overview
Execute these phases in order, completing ALL of them:
- Discovery - Understand the feature requirements
- Codebase Exploration - Map relevant code areas
- Clarifying Questions - Resolve ambiguities
- Architecture Design - Design the implementation approach
- Implementation - Build the feature
- Quality Review - Review for issues
- Summary - Document accomplishments
Phase 1: Discovery
Goal: Understand what the user wants to build.
-
Analyze the feature description from the user's input:
- What is the core functionality?
- What are the expected inputs and outputs?
- Are there any constraints mentioned?
- What success criteria can you infer?
-
Summarize your understanding to the user. Ask the user to confirm if your understanding is correct before proceeding.
Phase 2: Codebase Exploration
Goal: Understand the relevant parts of the codebase.
-
Run deep-analysis workflow:
- Read
../deep-analysis/SKILL.md and follow its workflow
- Pass the feature description from Phase 1 as the analysis context
- This handles reconnaissance, team planning, approval (auto-approved when skill-invoked), team creation, parallel exploration (code-explorer agents), and synthesis (code-synthesizer agent)
- Note: Deep-analysis may return cached results if a valid exploration cache exists. In skill-invoked mode, cache hits are auto-accepted — this is expected behavior that avoids redundant exploration.
-
Present the synthesized analysis to the user.
Phase 3: Clarifying Questions
Goal: Resolve any ambiguities before designing.
-
Review the feature requirements and exploration findings.
-
Identify underspecified aspects:
- Edge cases not covered
- Technical decisions that could go multiple ways
- Integration points that need clarification
- Performance or scale requirements
-
Ask clarifying questions:
Ask the user to answer critical unknowns. Only ask questions that would significantly impact the implementation.
If no clarifying questions are needed, inform the user and proceed.
Phase 4: Architecture Design
Goal: Design the implementation approach.
-
Load skills for this phase:
- Read
../architecture-patterns/SKILL.md and apply its guidance
- Read
../language-patterns/SKILL.md and apply its guidance
- Read
../technical-diagrams/SKILL.md and apply its styling rules for any Mermaid diagrams in architecture proposals
-
Launch code-architect subagents:
Invoke the code-architecture skill by reading ../code-architecture/SKILL.md and following its workflow. Spawn 2-3 instances with different approaches:
Agent 1: Design a minimal, focused approach prioritizing simplicity
Agent 2: Design a flexible, extensible approach prioritizing future changes
Agent 3: Design an approach optimized for the project's existing patterns (if applicable)
For each approach, invoke the code-architecture skill with the specific design brief:
Feature: [feature description]
Design approach: [specific approach for this agent]
Based on the codebase exploration:
[Summary of relevant files and patterns]
Design an implementation that:
- Lists files to create/modify
- Describes the changes needed in each file
- Explains the data flow
- Identifies risks and mitigations
Return a detailed implementation blueprint.
-
Present approaches:
- Summarize each approach
- Compare trade-offs (simplicity, flexibility, performance, maintainability)
- Make a recommendation with justification
-
User chooses approach:
Ask the user to select an approach or request modifications.
-
Generate ADR artifact:
- Read the ADR template from
references/adr-template.md
- Create an ADR documenting:
- Context: Why this feature is needed
- Decision: The chosen approach
- Consequences: Trade-offs and implications
- Alternatives: Other approaches considered
- Determine the next ADR number by checking existing files in
internal/docs/adr/
- Save to
internal/docs/adr/NNNN-[feature-slug].md (create internal/docs/adr/ if needed)
- Inform the user of the saved ADR location
Phase 5: Implementation
Goal: Build the feature.
-
Require explicit approval:
Ask the user: "Ready to begin implementation of [feature] using [chosen approach]?"
Wait for confirmation before proceeding.
-
Read all relevant files:
Before making any changes, read the complete content of every file you'll modify.
-
Implement the feature:
- Follow the chosen architecture design
- Match existing code patterns and conventions
- Create new files as needed
- Update existing files using Edit tool
- Add appropriate error handling
- Include inline comments only where logic isn't obvious
-
Test if applicable:
- If the project has tests, add tests for the new functionality
- Run existing tests to ensure nothing broke
-
IMPORTANT: Proceed immediately to Phase 6.
Do NOT stop here. Do NOT wait for user input. Implementation is complete, but the workflow requires Quality Review and Summary phases. Continue directly to Phase 6 now.
Phase 6: Quality Review
Goal: Review the implementation for issues.
-
Load skills for this phase:
- Read
../code-quality/SKILL.md and apply its guidance
-
Launch code-reviewer subagents:
Spawn 3 subagents using the code-reviewer instructions from agents/code-reviewer.md with different focuses:
Agent 1: Review for correctness and edge cases
Agent 2: Review for security and error handling
Agent 3: Review for maintainability and code quality
For each review focus, spawn a subagent with the code-reviewer instructions from agents/code-reviewer.md and the specific review brief:
Review focus: [specific focus for this agent]
Files to review:
[List of files modified/created]
Review the implementation and report:
- Issues found with confidence scores (0-100)
- Suggestions for improvement
- Positive observations
Only report issues with confidence >= 80.
-
Aggregate findings:
- Collect results from all reviewers
- Deduplicate similar issues
- Prioritize by severity and confidence
-
Present findings:
Show the user:
- Critical issues (must fix)
- Moderate issues (should fix)
- Minor suggestions (nice to have)
-
User decides:
Ask the user:
- "Fix all issues now"
- "Fix critical issues only"
- "Proceed without fixes"
- "I'll fix manually later"
-
If fixing: make the changes and re-review if needed.
-
IMPORTANT: Proceed immediately to Phase 7.
Do NOT stop here. The workflow requires a Summary phase to document accomplishments and update the CHANGELOG. Continue directly to Phase 7 now.
Phase 7: Summary
Goal: Document and celebrate accomplishments.
-
Summarize accomplishments:
Present to the user:
- What was built
- Key files created/modified
- Architecture decisions made
- Architecture diagram (Mermaid flowchart showing the implemented structure)
- Any known limitations or future work
-
Update CHANGELOG.md:
- Read the entry template from
references/changelog-entry-template.md
- Load the
changelog-format skill for Keep a Changelog guidelines
- Create an entry under the
[Unreleased] section with:
- Appropriate category (Added, Changed, Fixed, etc.)
- Concise description of the feature
- If
CHANGELOG.md doesn't exist, create it with proper header
- Add the entry to the appropriate section under
[Unreleased]
- Inform the user of the update
-
Final message:
Congratulate the user and offer next steps:
- Commit the changes
- Create a PR
- Additional testing suggestions
Error Handling
If any phase fails:
- Explain what went wrong
- Ask the user how to proceed:
- Retry the phase
- Skip to next phase
- Abort the workflow
Agents
This skill uses the following agents directly:
| Agent | File | Dependencies |
|---|
| code-reviewer | agents/code-reviewer.md | none |
Additionally, this skill invokes the following skills for agent access:
deep-analysis — for codebase exploration and synthesis (Phase 2)
code-architecture — for architectural design via code-architect agents (Phase 4)
Execution Strategy
Execute agents respecting their dependency graph.
If subagent dispatch is available: Dispatch code-architecture invocations in parallel for different design approaches (Phase 4). Dispatch code-reviewer agents in parallel with different review focuses (Phase 6), passing the contents of agents/code-reviewer.md as the task instructions. Wait for all subagents to complete before proceeding.
If subagent dispatch is not available: For Phase 4, invoke the code-architecture skill sequentially for each design approach. For Phase 6, read agents/code-reviewer.md and follow its instructions sequentially for each review focus. Write outputs before proceeding to the next review.
Agent Coordination
- Exploration and synthesis are handled by the
deep-analysis skill in Phase 2 (auto-approved when skill-invoked)
- Architectural design is handled by the
code-architecture skill in Phase 4
- Code review uses this skill's own
agents/code-reviewer.md in Phase 6
- Give each subagent a distinct focus area
- Wait for all subagents to complete before proceeding
- Handle subagent failures gracefully (continue with partial results)
- Use a high-capability model for architecture design and code review
Additional Resources