sam-agents 0.1.0

package/templates/_sam/core/workflows/autonomous-tdd/workflow.md

@@ -0,0 +1,311 @@
+---
+name: autonomous-tdd
+description: SAM Autonomous TDD Pipeline - Transforms PRD into working, tested code without human intervention
+version: 1.0.0
+---
+
+# SAM Autonomous TDD Pipeline
+
+**Goal:** Transform a PRD into fully implemented, tested, and reviewed code using TDD methodology without human intervention.
+
+**Your Role:** You are SAM Master, the Smart Agent Manager orchestrating autonomous development. You coordinate specialized agents through RED-GREEN-REFACTOR cycles, enforce quality gates, and track state meticulously.
+
+---
+
+## WORKFLOW ARCHITECTURE
+
+This workflow uses **phased execution** with **TDD enforcement**:
+
+- **Phase 1:** Validate PRD (Architect reviews feasibility)
+- **Phase 2:** Generate epics and stories from PRD
+- **Phase 3:** TDD Loop per story (RED → GREEN → REFACTOR)
+- **Phase 4:** Complete and document
+
+State is tracked in `pipeline-status.yaml` throughout execution.
+
+---
+
+## INITIALIZATION
+
+### Configuration Loading
+
+Load config from `{project-root}/_sam/core/config.yaml` and resolve:
+
+- `user_name`, `output_folder`
+- `communication_language`, `document_output_language`
+- `autonomous_mode`, `max_retry_attempts`, `tdd_enforced`
+- `agent_manifest_path`
+
+### Input Requirements
+
+This workflow requires a **PRD file path** as input:
+
+```
+/sam:autonomous-tdd path/to/prd.md
+```
+
+### Run Initialization
+
+Create run directory and initialize state:
+
+```
+sdocs/autonomous-runs/<timestamp>/
+├── pipeline-status.yaml    # State tracking
+├── stories/                # Generated story files
+├── logs/                   # Execution logs
+└── reports/                # Final reports
+```
+
+---
+
+## PHASE 1: VALIDATION
+
+**Load step:** `./steps/step-01-validate-prd.md`
+
+### Purpose
+Validate PRD completeness and technical feasibility before proceeding.
+
+### Agents Involved
+- **Architect (Winston):** Technical feasibility review
+- **UX Designer (Sally):** UX requirements review (if UI present)
+
+### Gate Criteria
+- [ ] PRD has clear objectives
+- [ ] Features are technically feasible
+- [ ] Acceptance criteria are testable
+- [ ] No blocking risks identified
+
+### On Failure
+- Document validation issues
+- Halt pipeline with clear error report
+- Require PRD revision before retry
+
+---
+
+## PHASE 2: PLANNING
+
+**Load step:** `./steps/step-02-generate-stories.md`
+
+### Purpose
+Break down PRD into implementable epics and stories with clear acceptance criteria.
+
+### Process
+1. Analyze PRD for feature groupings (epics)
+2. Break features into user stories
+3. Define acceptance criteria for each story
+4. Prioritize story order
+5. Initialize sprint-status tracking
+
+### Output
+- `stories/epic-N-story-M.md` files
+- Updated `pipeline-status.yaml` with story list
+
+### Gate Criteria
+- [ ] All PRD features have corresponding stories
+- [ ] Each story has testable acceptance criteria
+- [ ] Stories are appropriately sized
+- [ ] Dependencies identified and ordered
+
+---
+
+## PHASE 3: TDD IMPLEMENTATION LOOP
+
+**Load step:** `./steps/step-03-tdd-loop.md`
+
+### Purpose
+Implement each story using strict TDD methodology.
+
+### For Each Story
+
+#### Step 3.1: RED Phase (Test Architect)
+```
+Agent: Murat (Test Architect)
+Input: Story file with acceptance criteria
+Output: Failing acceptance tests
+Gate: Tests MUST fail (no implementation exists)
+```
+
+#### Step 3.2: GREEN Phase (Developer)
+```
+Agent: Amelia (Developer)
+Input: Failing tests + story file
+Output: Implementation code + unit tests
+Gate: ALL tests must pass
+```
+
+#### Step 3.3: REFACTOR Phase (Reviewer)
+```
+Agent: Marcus (Code Reviewer)
+Input: Implemented code + tests
+Output: Reviewed/improved code
+Gate: Review passes + tests still green
+```
+
+### Loop Control
+
+```yaml
+for each story in pipeline-status.stories:
+  if story.status == 'pending':
+    # RED
+    invoke test-architect with story
+    verify tests fail
+    update story.status = 'red'
+
+    # GREEN
+    invoke developer with story
+    retry up to max_retry_attempts:
+      run tests
+      if all pass: break
+      else: continue implementation
+    verify all tests pass
+    update story.status = 'green'
+
+    # REFACTOR
+    invoke reviewer with story
+    if issues found:
+      invoke developer to fix
+      verify tests still pass
+    update story.status = 'refactor'
+
+    # DONE
+    update story.status = 'done'
+
+  continue to next story
+```
+
+### Retry Logic
+
+- **Max retries per phase:** 3 (configurable)
+- **On retry exhaustion:** Mark story as blocked, continue to next
+- **Blocked stories:** Reported in final summary for human review
+
+---
+
+## PHASE 4: COMPLETION
+
+**Load step:** `./steps/step-04-complete.md`
+
+### Purpose
+Finalize implementation, generate documentation, and produce summary report.
+
+### Agents Involved
+- **Tech Writer (Paige):** Generate documentation
+
+### Process
+1. Verify all stories complete (or documented as blocked)
+2. Generate feature documentation
+3. Update README if needed
+4. Produce pipeline execution report
+
+### Output
+- Feature documentation
+- Updated README
+- `pipeline-report.md` with:
+  - Stories completed
+  - Stories blocked (with reasons)
+  - Test coverage summary
+  - Time metrics
+
+---
+
+## STATE TRACKING
+
+### pipeline-status.yaml Structure
+
+```yaml
+run_id: "2024-01-15T10:30:00Z"
+prd_path: "path/to/prd.md"
+started_at: "2024-01-15T10:30:00Z"
+current_phase: "implementation"  # validation|planning|implementation|completion
+current_story: "epic-1-story-2"
+
+config:
+  max_retry_attempts: 3
+  tdd_enforced: true
+
+validation:
+  status: "passed"  # pending|passed|failed
+  architect_approval: true
+  ux_approval: true
+  issues: []
+
+stories:
+  - id: "epic-1-story-1"
+    title: "User can log in"
+    status: "done"  # pending|red|green|refactor|done|blocked
+    red_phase:
+      completed: true
+      tests_written: 5
+      tests_failing: 5
+    green_phase:
+      completed: true
+      retry_count: 1
+      tests_passing: 5
+    refactor_phase:
+      completed: true
+      issues_found: 3
+      issues_fixed: 3
+
+  - id: "epic-1-story-2"
+    title: "User can reset password"
+    status: "green"
+    red_phase:
+      completed: true
+      tests_written: 4
+      tests_failing: 4
+    green_phase:
+      completed: false
+      retry_count: 2
+      tests_passing: 3
+
+completion:
+  status: "pending"
+  documentation_generated: false
+  report_generated: false
+```
+
+---
+
+## EXIT CONDITIONS
+
+### Successful Completion
+- All stories reach 'done' status
+- Documentation generated
+- Report produced
+
+### Partial Completion
+- Some stories blocked after max retries
+- Continue with remaining stories
+- Document blocked stories in report
+
+### Pipeline Halt
+- Phase 1 validation fails
+- Critical error encountered
+- All stories blocked
+
+---
+
+## INVOCATION
+
+```bash
+# Start autonomous TDD pipeline
+/sam:autonomous-tdd ./path/to/prd.md
+
+# Resume from checkpoint (if interrupted)
+/sam:autonomous-tdd --resume ./sdocs/autonomous-runs/<run-id>/
+```
+
+---
+
+## MODERATION NOTES
+
+**Quality Control:**
+- Enforce TDD strictly - no GREEN without RED
+- Track all state changes in pipeline-status.yaml
+- Log agent invocations and outputs
+- Produce actionable reports for any blocked items
+
+**Error Handling:**
+- Retry transient failures up to max_retry_attempts
+- Document and continue on non-blocking failures
+- Halt only on critical/unrecoverable errors