The workflow execution engine is governed by: {project-root}/_bmad/core/tasks/workflow.xml
You MUST have already loaded and processed: {installed_path}/workflow.yaml
Communicate all responses in {communication_language} and language MUST be tailored to {user_skill_level}
Generate all documents in {document_output_language}
๐ SUPER-DEV MODE: Enhanced quality workflow with post-implementation validation and automated code review
This workflow ensures stories are TRULY complete through multi-stage validation before marking done
โ๏ธ EXECUTION FLOW: First execute ALL dev-story steps (1-8), then continue with super-dev enhancements (9.5-9.6)
๐ฏ EXECUTE DEV-STORY STEPS 1-8 FIRST
Follow all instructions in: {project-root}/_bmad/bmm/workflows/4-implementation/dev-story/instructions.xml
Complete Steps 1 through 8 exactly as specified in dev-story workflow
After Step 8 completes (all tasks checked), return here and continue with Step 9.5 below
โ๏ธ Super-dev-story includes ALL standard dev-story steps (1-8):
**Standard Development Flow:**
- Step 1: Find and load story (sprint-status or user-provided path)
- Step 1.5: โ
PRE-DEV GAP ANALYSIS - Validate tasks against codebase
- Step 2: Load project context and story information
- Step 3: Detect review continuation (if resuming after code-review)
- Step 4: Mark story in-progress in sprint-status
- Step 5: Implement task following red-green-refactor cycle (TDD)
- Step 6: Author comprehensive tests (unit, integration, e2e)
- Step 7: Run validations and tests (regression suite)
- Step 8: Validate and mark task complete (strict gates)
After completing Steps 1-8, super-dev mode adds ADDITIONAL quality validation below.
**For complete details of Steps 1-8, see:** {project-root}/_bmad/bmm/workflows/4-implementation/dev-story/instructions.xml
๐ POST-DEV VALIDATION - Verify all work actually completed!
This catches incomplete implementations that were prematurely marked done
Re-read story requirements and acceptance criteria
Extract all tasks and subtasks that are marked [x] complete
For each completed task, identify what should exist in codebase
Use Glob to find files that should have been created
Use Grep to search for functions/classes that should exist
Use Read to verify implementation completeness (not just existence)
Run tests to verify claimed test coverage actually exists and passes
Compare claimed work vs actual implementation:
**POST-DEV VERIFICATION:**
โ
Verified Complete:
- List tasks where code fully exists and works
- Confirm tests exist and pass
- Verify implementation matches requirements
โ False Positives Detected:
- List tasks marked [x] but code missing or incomplete
- Identify claimed tests that don't exist or fail
- Note partial implementations marked as complete
โ ๏ธ Quality Issues:
- Implementations that exist but don't meet requirements
- Tests that exist but have poor coverage
- Code that works but violates architecture patterns
Uncheck false positive tasks in story file
Add new tasks for missing work to Tasks/Subtasks section
Add note to Dev Agent Record: "Post-dev gap analysis found incomplete work - continuing implementation"
Update Gap Analysis section with post-dev findings
Set {{fix_iteration_count}} = {{fix_iteration_count}} + 1
Add to Dev Agent Record: "HALTED - exceeded max fix iterations, manual review needed"
HALT - require user intervention
Continue implementation with added tasks
Update Gap Analysis section with post-dev verification results
Continue to Step 9.6
๐ AUTO CODE REVIEW - Independent quality validation
Fresh perspective catches issues the dev agent might miss
Extract File List from Dev Agent Record
Identify all files created or modified during implementation
Perform comprehensive code review checking:
- **Correctness:** Logic errors, edge cases, error handling
- **Architecture:** Compliance with patterns and standards
- **Security:** Vulnerabilities, input validation, authentication
- **Performance:** Inefficiencies, N+1 queries, memory leaks
- **Testing:** Test coverage, edge cases, test quality
- **Code Quality:** Readability, maintainability, documentation
Categorize all findings by severity:
- CRITICAL: Security vulnerabilities, data loss, broken functionality
- HIGH: Logic errors, missing error handling, test gaps
- MEDIUM: Code quality issues, minor bugs, performance concerns
- LOW: Style issues, documentation improvements, minor refactoring
Add "Code Review Findings" section to story file
Add review findings as new tasks in Tasks/Subtasks with [AI-Review] prefix
Add to Dev Agent Record: "Code review found {{total_issue_count}} issues - continuing with fixes"
Set {{fix_iteration_count}} = {{fix_iteration_count}} + 1
Add to Dev Agent Record: "HALTED - code review issues after max iterations, manual review needed"
HALT - require user intervention
Implement fixes
Auto-fix these minor issues?
Options:
[Y] Yes - Add to task list and fix now
[n] No - Document in story but don't fix (can address later)
[s] Skip - Ignore these findings
Add review findings as tasks with [AI-Review] prefix
Add to Dev Agent Record: "Addressing {{issue_count}} minor code review findings"
Implement fixes
Add "Code Review Findings" section documenting issues
Add to Dev Agent Record: "Code review found {{issue_count}} minor issues - documented for future work"
Continue to Step 9
Continue to Step 9
Add to Dev Agent Record: "Code review completed - no issues found"
Continue to Step 9
๐ฏ FINAL VALIDATION - All quality gates must pass!
Verify ALL tasks and subtasks are marked [x] (re-scan the story document now)
Run the full regression suite (do not skip)
Confirm File List includes every changed file
Execute enhanced definition-of-done validation
Validate super-dev quality gates:
- โ
Pre-dev gap analysis completed and approved
- โ
All tasks/subtasks marked complete with [x]
- โ
Implementation satisfies every Acceptance Criterion
- โ
All tests pass (no regressions, new tests successful)
- โ
Post-dev gap analysis passed (no false positives)
- โ
Code review completed (critical/high issues resolved)
- โ
File List includes every new/modified/deleted file
- โ
Dev Agent Record contains implementation notes
- โ
Change Log includes summary of changes
Update the story Status to: "review"
Load the FULL file: {sprint_status}
Find development_status key matching {{story_key}}
Verify current status is "in-progress" (expected previous state)
Update development_status[{{story_key}}] = "review"
Save file, preserving ALL comments and structure including STATUS DEFINITIONS
HALT - Complete remaining tasks before marking ready for review
HALT - Fix regression issues before completing
HALT - Update File List with all changed files
HALT - Run post-dev validation before completing
HALT - Run code review before completing
Prepare comprehensive summary in Dev Agent Record โ Completion Notes
Based on {user_skill_level}, ask if user needs any explanations about:
- What was implemented and how it works
- Why certain technical decisions were made
- Gap analysis or code review findings
- Any patterns, libraries, or approaches used
Provide clear, contextual explanations tailored to {user_skill_level}
Use examples and references to specific code when helpful
๐ PUSH-ALL - Stage, commit, and push all changes with safety validation
โ ๏ธ Use with caution - commits ALL changes in repository
Run git commands in parallel:
- git status - Show modified/added/deleted/untracked files
- git diff --stat - Show change statistics
- git log -1 --oneline - Show recent commit for message style
- git branch --show-current - Confirm current branch
Parse git status output to identify:
- Modified files
- Added files
- Deleted files
- Untracked files
- Total insertion/deletion counts
๐ SAFETY VALIDATION - Check for secrets, large files, and unwanted artifacts
Scan all changed files for dangerous patterns:
**Secret Detection:**
Check for files matching secret patterns:
- .env*, *.key, *.pem, credentials.json, secrets.yaml
- id_rsa, *.p12, *.pfx, *.cer
- Any file containing: _API_KEY=, _SECRET=, _TOKEN= with real values
Validate API keys are placeholders only:
โ
Acceptable: API_KEY=your-api-key-here, SECRET=placeholder, TOKEN=xxx, API_KEY=${'{YOUR_KEY}'}
โ BLOCK: OPENAI_API_KEY=sk-proj-*, AWS_SECRET_KEY=AKIA*, STRIPE_API_KEY=sk_live_*
**File Size Check:**
Check for files >10MB without Git LFS
**Build Artifacts:**
Check for unwanted directories/files:
- node_modules/, dist/, build/, __pycache__/, *.pyc, .venv/
- .DS_Store, thumbs.db, *.swp, *.tmp
**Git State:**
Verify:
- .gitignore properly configured
- No merge conflicts
- Repository initialized
HALT - Cannot proceed with secrets in changeset
Proceed with large files? [y/n]:
HALT
Proceed despite build artifacts? [y/n]:
HALT
Push directly to {{branch_name}}? [y/n]:
HALT
**Proceed with commit and push?**
Type 'yes' to proceed, 'no' to cancel, or 'review' to see detailed diff:
Run: git diff --stat
Show detailed file changes
Still proceed with commit and push? [yes/no]:
HALT - User cancelled
Execute: git add .
Execute: git status
Analyze changes to determine commit type and scope:
- feat: New feature implementation
- fix: Bug fixes
- docs: Documentation updates
- refactor: Code restructuring
- test: Test additions/updates
- chore: Tooling, configs, dependencies
Generate commit message in conventional commit format:
Format:
[type](optional scope): Brief summary (max 72 chars)
- Key change 1
- Key change 2
- Key change 3
Base message on:
- Story title and key
- Files changed
- Acceptance criteria met
- Recent commit style
Use this commit message? [yes/edit/cancel]:
Enter your commit message:
Use user-provided message as {{commit_message}}
HALT
Set {{commit_message}} = {{generated_commit_message}}
Execute git commit with message:
git commit -m "$(cat <<'EOF'
{{commit_message}}
EOF
)"
HALT - Fix errors before proceeding
Execute: git push
Execute: git pull --rebase
Execute: git push
Execute: git push -u origin {{current_branch}}
HALT - Manual push required
Execute: git log -1 --oneline --decorate