workflow-presets

Instruction file imported from wtthornton/TappsCodingAgents (.cursor/rules/workflow-presets.mdc). Copyright stays with the author.

Workflow Presets - Quick Commands

Overview

TappsCodingAgents provides preset workflows that can be executed with simple one-word commands. These workflows automatically orchestrate multiple agents to complete common SDLC tasks.

NEW (v3.5.37+): All workflows now include adaptive checkpoints that optimize execution:

• Checkpoint 1 (After Enhance): Early validation using prompt analysis
• Checkpoint 2 (After Planning): Task complexity analysis with workflow switching
• Checkpoint 3 (After Test): Quality-based early termination

Workflows automatically adapt to task complexity, potentially using fewer steps than listed. See Checkpoint System Guide for details.

Available Workflow Presets

1. Full SDLC Pipeline (full / enterprise)

Command: python -m tapps_agents.cli workflow full

Best for: Complete software development lifecycle with all quality gates

Typical Steps: 5-10 (adaptive based on checkpoints)

Agent Sequence (with checkpoints):

1. Enhancer - Enhance Prompt ✅ Checkpoint 1: Early mismatch detection (may switch workflow)
2. Analyst - Gather Requirements
3. Planner - Create Stories ✅ Checkpoint 2: Task complexity analysis (may switch to *build or *fix)
4. Architect - Design System
5. Designer - Api Design
6. Implementer - Write Code
7. Reviewer - Review Code
8. Tester - Write Tests ✅ Checkpoint 3: Quality gate (may skip steps 9-10)
9. Ops - Security Scan (optional if quality ≥ 80)
10. Documenter - Generate Docs (optional if quality ≥ 75)

Quality Gates:

• Overall score: ≥70
• Security: ≥7.0
• Maintainability: ≥7.0

Checkpoint Optimization:

• May switch to *build if task complexity is medium (saves 20K-40K tokens)
• May switch to *fix if task is simple bug fix (saves 30K-50K tokens)
• May skip security + docs if quality score ≥ 80 (saves 12K-13K tokens)

2. Fix (fix / maintenance / hotfix / urgent / refactor / quick-fix)

Command: python -m tapps_agents.cli workflow fix

Best for: Debug, fix, and verify issues in existing code

Typical Steps: 3-6 (adaptive based on checkpoints)

Agent Sequence (with checkpoints):

1. Enhancer - Enhance Prompt ✅ Checkpoint 1: Early validation (confirms this is a fix, not a feature)
2. Debugger - Analyze Error
3. Implementer - Write Code
4. Reviewer - Review Code
5. Tester - Write Tests ✅ Checkpoint 3: Quality gate (may skip step 6 if quality ≥ 75)
6. Orchestrator - Finalize (optional based on quality)

Quality Gates:

• Overall score: ≥65
• Security: ≥6.5

Checkpoint Optimization:

• Checkpoint 1 may recommend staying with *fix for clear bug fixes
• Checkpoint 3 may skip finalization steps if quality is excellent

3. Quality Improvement (quality / improve)

Command: python -m tapps_agents.cli workflow quality

Best for: Comprehensive code review and improvement cycle

Agent Sequence:

1. Reviewer - Review Code
2. Improver - Refactor
3. Reviewer - Review Code
4. Tester - Write Tests
5. Ops - Security Scan
6. Orchestrator - Finalize

Quality Gates:

• Overall score: ≥75
• Security: ≥7.5
• Maintainability: ≥8.0

4. Brownfield Project Analysis (brownfield)

Command: python -m tapps_agents.cli workflow brownfield

Best for: Comprehensive analysis workflow for existing codebases that provides:

• Project discovery and profiling
• Code quality assessment
• Architecture pattern analysis
• Codebase structure evaluation
• Technical debt categorization and prioritization
• Dependency and integration mapping
• Test and documentation coverage analysis
• Modernization roadmap generation

This workflow orchestrates multiple agents to provide a complete picture of a brownfield project's current state and recommended improvements.

Agent Sequence:

Usage in Cursor AI

When working in Cursor AI, you can:

1. Run workflows directly:

   • Type: "Run rapid development workflow"
   • Or: "Execute full SDLC pipeline"
   • Or: "Start quality improvement"

2. Use voice commands:

   • "run rapid dev"
   • "execute enterprise workflow"
   • "start quick fix"

3. Reference in conversations:

   • "Use the rapid workflow for this feature"
   • "This needs the full enterprise workflow"
   • "Run the quality improvement cycle"

When to Use Each Workflow

• Complete software development lifecycle with all quality gates → workflow full
• Debug, fix, and verify issues in existing code → workflow fix
• Comprehensive code review and improvement cycle → workflow quality
• **Comprehensive analysis workflow for existing codebases that provides:
• Project discovery and profiling
• Code quality assessment
• Architecture pattern analysis
• Codebase structure evaluation
• Technical debt categorization and prioritization
• Dependency and integration mapping
• Test and documentation coverage analysis
• Modernization roadmap generation

This workflow orchestrates multiple agents to provide a complete picture of a brownfield project's current state and recommended improvements** → workflow brownfield

Integration

These workflows integrate with:

• ✅ Expert consultation (domain experts via @expert *consult)
• ✅ Quality gates (scoring thresholds)
• ✅ Context tiers (token optimization)
• ✅ Multi-agent orchestration
• ✅ Git worktree isolation
• ✅ Epic orchestration (story-only or full workflow per story)
• ✅ Parallel story execution (wave-based with dependency ordering)

List All Presets

python -m tapps_agents.cli workflow list

Customization

Workflow presets are defined in workflows/presets/:

• full-sdlc.yaml
• fix.yaml
• quality.yaml
• brownfield-analysis.yaml

You can modify these files or create your own presets.