@zeyue0329/xiaoma-cli 1.11.0 → 1.13.0

package/src/xmc/workflows/5-full-pipeline/auto-full-pipeline/steps/step-04-run-story-pipeline.md

@@ -1,14 +1,14 @@
 ---
 name: "step-04-run-story-pipeline"
-description: "Execute the complete auto-story-pipeline in batch mode to develop, review, test, and deliver all stories"
+description: "Execute the auto-story-pipeline-batch to develop, review, test, and deliver all stories using Agent subprocess isolation"
 nextStepFile: "./step-05-finalize.md"
 ---
 
-# Step 4 of 5: Execute Story Development Pipeline
+# Step 4 of 5: Execute Story Development Pipeline (Batch)
 
-**Goal:** Delegate to the auto-story-pipeline in **batch mode** to execute the complete story development lifecycle for ALL stories — from story creation through validation, development, code review, testing, bug fixing, and final delivery.
+**Goal:** Delegate to the **auto-story-pipeline-batch** to execute the complete story development lifecycle for ALL stories. Each story is processed in an independent Agent subprocess for context isolation.
 
-**Role:** Master Pipeline Orchestrator → delegates to auto-story-pipeline roles (SM xiaomin → PM xiaochan → Dev xiaokai → Reviewer → QA xiaoce)
+**Role:** Master Pipeline Orchestrator → delegates to auto-story-pipeline-batch scheduler
 
 ---
 
@@ -19,14 +19,15 @@ nextStepFile: "./step-05-finalize.md"
 **Output:**
 
 ```
-═══════════════════════════════════════════════════
+===============================================
   Phase 3 of 3 — Story Development Pipeline
-═══════════════════════════════════════════════════
+===============================================
 
 Starting automated story development in BATCH mode...
+Using Agent subprocess isolation — each story processed independently.
 Processing all {backlog_count} backlog stories sequentially.
 
-For each story, the pipeline will:
+For each story, an independent Agent will:
   1. Create story file from epics
   2. Validate story (PM quality gate)
   3. Develop implementation (TDD)
@@ -35,37 +36,27 @@ For each story, the pipeline will:
   6. Fix-and-retest loop (if needed)
   7. Mark story as done
 
-Delegating to auto-story-pipeline...
-───────────────────────────────────────────────────
+Delegating to auto-story-pipeline-batch...
+-----------------------------------------------
 ```
 
-### 2. Execute Auto Story Pipeline
+### 2. Execute Auto Story Pipeline Batch
 
-Read fully and follow: `{project-root}/_xiaoma/xmc/workflows/4-implementation/auto-story-pipeline/workflow.md`
+Read fully and follow: `{project-root}/_xiaoma/xmc/workflows/4-implementation/auto-story-pipeline-batch/workflow.md`
 
 **Critical Full Pipeline Mode Instructions:**
 
 - **FULLY AUTOMATED:** Do NOT pause for user input at any point
-- **BATCH MODE:** When step-01 of the story pipeline asks for pipeline mode, select "batch" — process ALL backlog stories sequentially
-- Execute ALL 9 sub-steps of the story pipeline (step-01 through step-09)
-- The story pipeline will loop through its step-02 → step-09 cycle for each story:
-  - step-02: Create user story from epics and planning artifacts
-  - step-03: PM validation (quality gate, max 3 attempts)
-  - step-04: Development implementation (TDD)
-  - step-05: Adversarial code review (auto-fix HIGH/MEDIUM issues)
-  - step-06: QA functional testing (real data, no mocks)
-  - step-07: Fix-and-retest loop (max `{max_fix_iterations}` iterations)
-  - step-08: Complete story (mark done, update sprint status)
-  - step-09: Cycle check (find next story, loop back to step-02 or finalize)
-- **IMPORTANT:** The story pipeline's step-01 will validate prerequisites including sprint-status.yaml — since we just generated it in step-03 of this pipeline, this check should pass
-- **IMPORTANT:** The story pipeline's step-01 may attempt to auto-generate sprint-status.yaml if it can't find it — this is fine as a safety net, but should not be needed
-- Continue through ALL stories until the story pipeline reaches finalization in step-09 and completes
-- **Do NOT stop after individual story completions** — the batch mode cycle check (step-09) handles continuation automatically
-- **If a story HALTS** (e.g., fix iteration exceeded): The story pipeline's escalation logic handles this. The escalation report will be generated. The full pipeline should note this as a partial completion
+- The batch pipeline will automatically process ALL backlog and ready-for-dev stories
+- Each story is handled by an independent Agent subprocess (isolated context)
+- The batch scheduler stays lightweight — only orchestrating the loop
+- **Do NOT stop after individual story completions** — the batch scheduler handles continuation automatically
+- **If a story Agent fails:** The batch scheduler records the failure and continues to the next story
+- Continue until all stories are processed or the batch pipeline finalizes
 
 ### 3. Verify Story Pipeline Output
 
-After the story pipeline completes (either all stories done or a HALT condition):
+After the batch pipeline completes:
 
 1. **Re-read sprint-status.yaml** — Fresh read from disk to get final state
 2. **Count final statuses:**
@@ -85,24 +76,24 @@ After the story pipeline completes (either all stories done or a HALT condition)
 **Output:**
 
 ```
-───────────────────────────────────────────────────
-  ✅/⚠️/❌ Phase 3 Complete — Story Development
-───────────────────────────────────────────────────
+-----------------------------------------------
+  Phase 3 Complete — Story Development
+-----------------------------------------------
 
 Stories Completed: {done_count}/{total_story_count}
 Stories Remaining: {remaining_count}
 
 Sprint Status Breakdown:
-  ✅ Done: {done_count}
-  ⬜ Backlog: {backlog_count}
-  🔄 In Progress: {in_progress_count}
-  👀 Review: {review_count}
+  Done: {done_count}
+  Backlog: {backlog_count}
+  In Progress: {in_progress_count}
+  Review: {review_count}
 
 Phase 3 Status: {phase_3_status}
 Master Step: 4/5
 
 Proceeding to Finalization...
-───────────────────────────────────────────────────
+-----------------------------------------------
 ```
 
 **Auto-Proceed:** YES — Do NOT wait for user input. Immediately proceed to step-05.
@@ -111,15 +102,13 @@ Proceeding to Finalization...
 
 ## SUCCESS METRICS
 
-- Story pipeline executed in batch mode
+- Batch story pipeline executed with Agent subprocess isolation
 - All stories processed through create → validate → develop → review → test → complete
 - All story statuses updated to "done" in sprint-status.yaml
 - Epic statuses updated where all stories complete
 
 ## FAILURE MODES
 
-- Story pipeline HALT propagated (escalation from fix loop)
-- Individual story validation fails after 3 attempts
-- Story creation fails (malformed epics)
-- Story pipeline unable to find backlog stories
+- Batch pipeline unable to find backlog stories
+- Individual story Agent failures (recorded but non-blocking)
 - Partial completion (some stories done, others stuck)

package/src/xmc/workflows/5-full-pipeline/auto-full-pipeline/steps/step-05-finalize.md

@@ -156,7 +156,7 @@ Write to: `{implementation_artifacts}/full-pipeline-status.json`
 
 🔜 Recommended Next Steps:
   1. Review escalation reports for blocked stories
-  2. Run /xiaoma-auto-story-pipeline to resume remaining stories
+  2. Run /xiaoma-auto-story-pipeline-batch to resume remaining stories
   3. Use /xiaoma-correct-course (CC) if scope changes are needed
 ```
 
@@ -168,7 +168,7 @@ Write to: `{implementation_artifacts}/full-pipeline-status.json`
 🔜 Recommended Next Steps:
   1. Check story pipeline logs for error details
   2. Verify epics and architecture are well-formed
-  3. Run /xiaoma-auto-story-pipeline manually to diagnose
+  3. Run /xiaoma-auto-story-pipeline-batch manually to diagnose
 ```
 
 ### 7. Pipeline Complete

package/src/xmc/workflows/5-full-pipeline/auto-full-pipeline/workflow.md

@@ -13,7 +13,7 @@ description: "Fully automated requirements-to-delivery pipeline: requirements an
   - Analyst (xiaofen) → Architect (xiaojia) → PM (xiaochan) → Architect (xiaojia)
 - **Phase 2 — Bridge** (sprint planning)
   - SM (xiaomin) — Generate sprint tracking from epics
-- **Phase 3 — Story Development** (delegates to auto-story-pipeline)
+- **Phase 3 — Story Development** (delegates to auto-story-pipeline-batch)
   - SM (xiaomin) → PM (xiaochan) → Dev (xiaokai) → Reviewer → QA (xiaoce)
 
 - Communicate all responses in {communication_language} and generate all documents in {document_output_language}
@@ -38,7 +38,7 @@ This uses **step-file architecture** with delegation to existing sub-pipelines:
 - Step-01: Initialize and validate — unified prerequisite check
 - Step-02: Execute auto-requirements-pipeline — produces PRD, epics, architecture
 - Step-03: Bridge — auto-run sprint-planning to generate sprint-status.yaml
-- Step-04: Execute auto-story-pipeline — develop all stories in batch mode
+- Step-04: Execute auto-story-pipeline-batch — develop all stories using Agent subprocess isolation
 - Step-05: Finalize — unified completion report across both phases
 
 ### State Variables
@@ -59,13 +59,11 @@ This uses **step-file architecture** with delegation to existing sub-pipelines:
 - `{steps_completed}` — Requirements pipeline step counter
 - `{validation_iteration}`, `{analysis_iteration}`, `{architecture_iteration}` — Retry counters
 
-**Inherited from auto-story-pipeline (Phase 3):**
+**Inherited from auto-story-pipeline-batch (Phase 3):**
 
-- `{pipeline_mode}` — Always "batch" in full pipeline
-- `{current_story_key}`, `{current_story_path}` — Current story tracking
-- `{fix_iteration}`, `{max_fix_iterations}` — Bug fix loop control
-- `{stories_completed}` — Story completion counter
-- `{backlog_count}`, `{ready_count}`, `{in_progress_count}`, `{review_count}`, `{done_count}` — Status counts
+- Story processing is delegated to independent Agent subprocesses via auto-story-pipeline-batch
+- Each story runs in isolated context — see auto-story-pipeline-batch/workflow.md for details
+- `{backlog_count}`, `{ready_count}`, `{in_progress_count}`, `{review_count}`, `{done_count}` — Status counts (refreshed after batch completion)
 
 ### Pipeline Flow
 
@@ -96,7 +94,7 @@ Load config from `{project-root}/_xiaoma/xmc/config.yaml` and resolve:
 - `installed_path` = `{project-root}/_xiaoma/xmc/workflows/5-full-pipeline/auto-full-pipeline`
 - `requirements_pipeline` = `{project-root}/_xiaoma/xmc/workflows/1-analysis/auto-requirements-pipeline/workflow.md`
 - `sprint_planning_workflow` = `{project-root}/_xiaoma/xmc/workflows/4-implementation/xiaoma-sprint-planning/workflow.md`
-- `story_pipeline` = `{project-root}/_xiaoma/xmc/workflows/4-implementation/auto-story-pipeline/workflow.md`
+- `story_pipeline_batch` = `{project-root}/_xiaoma/xmc/workflows/4-implementation/auto-story-pipeline-batch/workflow.md`
 - `validation` = `{installed_path}/checklist.md`
 
 ### Context

package/src/xmc/workflows/5-full-pipeline/auto-prd-to-stories/SKILL.md

@@ -0,0 +1,6 @@
+---
+name: xiaoma-auto-prd-to-stories
+description: "Automated PRD-to-Stories pipeline (NO development): prd.md → epics → sprint-status → per-story detailed files. Stops before dev/review/QA so business architects can evaluate user-story quality. Use when the user says 'PRD 转故事', 'auto prd to stories', or 'APS'"
+---
+
+Follow the instructions in [workflow.md](workflow.md).

package/src/xmc/workflows/5-full-pipeline/auto-prd-to-stories/checklist.md

@@ -0,0 +1,43 @@
+# Auto PRD-to-Stories Pipeline Completion Checklist
+
+Use this checklist to validate the PRD-to-Stories pipeline completed successfully. This pipeline stops BEFORE development, so only planning + sprint-status + per-story files are in scope.
+
+## Phase 1 — Epics Creation Validation
+
+- [ ] **Input PRD Present** — `{planning_artifacts}/prd.md` exists and was readable at pipeline start
+- [ ] **Epics Created** — `{planning_artifacts}/epics.md` exists with full epic/story breakdown
+- [ ] **PRD Coverage** — Every major functional requirement in `prd.md` is covered by at least one epic
+- [ ] **Story Granularity** — Every epic contains at least one user story
+- [ ] **Acceptance Criteria Present** — Every story has BDD-formatted acceptance criteria (Given/When/Then)
+
+## Phase 2 — Sprint Planning Bridge Validation
+
+- [ ] **Sprint Status Generated** — `{implementation_artifacts}/sprint-status.yaml` exists with valid YAML
+- [ ] **All Epics Represented** — Every epic from `epics.md` has an `epic-N` entry in sprint status
+- [ ] **All Stories Represented** — Every story from `epics.md` has a `N-N-kebab-name` entry in sprint status
+
+## Phase 3 — Per-Story File Generation Validation
+
+- [ ] **One File Per Story** — Every story key in sprint-status.yaml has a corresponding `{implementation_artifacts}/{story_key}.md` file (excluding failed keys)
+- [ ] **Top-Level Sections Present** — Every story file contains: `Status`, `Story`, `Acceptance Criteria`, `Tasks / Subtasks`, `Dev Notes`
+- [ ] **Dev Notes Detailed-Design Subsections Present** — Every story file's `Dev Notes` contains these three mandatory subsections (per the step-04 Detailed Design Output Contract):
+  - [ ] `Functional Requirements Addressed` — lists FR/feature references from prd.md or epics.md
+  - [ ] `Data Model & Database Design` — concrete tables/fields/constraints/indexes OR an explicit `TBD by architect` line with reason
+  - [ ] `Implementation Logic` — concrete interface signatures / algorithm / state machine / error handling OR an explicit `TBD by architect` line with reason
+- [ ] **No Silent Omissions** — No story file contains a Dev Notes subsection that is only a heading with no body content
+- [ ] **Story Status Correct** — Every successfully created story file has `Status: ready-for-dev`
+- [ ] **No Stories Beyond ready-for-dev** — No story file shows status `in-progress`, `review`, or `done` (this pipeline never enters those states)
+- [ ] **Failed Stories Documented** — Any story key in `failed_story_keys` has its failure recorded in the final report with an error reason
+
+## Cross-Phase Traceability Validation
+
+- [ ] **FR Coverage Map Filled** — `epics.md` contains a `### FR Coverage Map` section with concrete `FR-N → Epic X / Story X.Y` mappings (not raw `{{placeholder}}`)
+- [ ] **FR Reverse Traceability** — Every functional requirement extracted from `prd.md` is referenced by at least one story file (verified by step-05's reverse-traceability check). Uncovered FRs are documented in Structural Warnings.
+
+## Pipeline-Level Validation
+
+- [ ] **All 5 Master Steps Executed** — Steps 1–5 completed without HALT (or HALT properly reported)
+- [ ] **No Dev/Review/Test Side Effects** — No code files modified outside `{planning_artifacts}` and `{implementation_artifacts}`
+- [ ] **Phase Boundary Summaries Output** — End-of-phase state snapshots logged for context-recovery
+- [ ] **Final Report Generated** — step-05 produced a unified completion report listing all artifacts and counts
+- [ ] **Artifact Consistency** — `sprint-status.yaml` story keys match story files on disk one-to-one (allowing for documented failures)

package/src/xmc/workflows/5-full-pipeline/auto-prd-to-stories/steps/step-01-init-and-validate.md

@@ -0,0 +1,155 @@
+---
+name: 'step-01-init-and-validate'
+description: 'Validate prd.md and sub-workflow paths, initialize PRD-to-Stories pipeline state variables, and prepare for execution'
+nextStepFile: './step-02-create-epics.md'
+---
+
+# Step 1 of 5: Initialize and Validate Environment
+
+**Goal:** Verify the input `prd.md` exists, all delegated sub-workflows are accessible, and initialize master pipeline state variables.
+
+**Role:** Pipeline Orchestrator
+
+---
+
+## EXECUTION SEQUENCE
+
+### 1. Load Configuration
+
+Read config from `{project-root}/_xiaoma/xmc/config.yaml` and resolve:
+
+- `project_name`, `user_name`
+- `communication_language`, `document_output_language`
+- `user_skill_level`
+- `planning_artifacts`, `implementation_artifacts`
+- `date` as system-generated current datetime
+
+**HALT if config not found:** "XiaoMa config not found at `{project-root}/_xiaoma/xmc/config.yaml`. Run `npx @zeyue0329/xiaoma-cli install` first."
+
+### 2. Validate Required Input
+
+**Input PRD** — `{planning_artifacts}/prd.md` must exist and be readable.
+
+- If missing: HALT — "No PRD document found at `{planning_artifacts}/prd.md`. This pipeline requires a pre-existing PRD as input. Either:
+  1. Place your PRD at `{planning_artifacts}/prd.md`, OR
+  2. Run the full auto-requirements-pipeline (AR) first to generate a PRD from a requirements document."
+
+- Read the first 20 lines of `prd.md` to extract document title and confirm it is non-empty. If empty: HALT — "prd.md exists but is empty."
+
+### 3. Validate Sub-Workflow Paths
+
+Check that all delegated sub-workflows are readable:
+
+1. **Create Epics & Stories** — `{project-root}/_xiaoma/xmc/workflows/3-solutioning/xiaoma-create-epics-and-stories/workflow.md`
+   - If missing: HALT — "xiaoma-create-epics-and-stories workflow not found. Verify XiaoMa installation is complete."
+
+2. **Sprint Planning** — `{project-root}/_xiaoma/xmc/workflows/4-implementation/xiaoma-sprint-planning/workflow.md`
+   - If missing: HALT — "xiaoma-sprint-planning workflow not found. Verify XiaoMa installation is complete."
+
+3. **Create Story** — `{project-root}/_xiaoma/xmc/workflows/4-implementation/xiaoma-create-story/workflow.md`
+   - If missing: HALT — "xiaoma-create-story workflow not found. Verify XiaoMa installation is complete."
+
+4. **Checklist** — `{installed_path}/checklist.md` must be readable for step-05 self-check.
+   - If missing: Output WARNING — "checklist.md not found; final self-check will be skipped."
+
+### 4. Check Existing Artifacts (Informational)
+
+Scan for pre-existing output artifacts from prior runs:
+
+- `{planning_artifacts}/epics.md`
+- `{implementation_artifacts}/sprint-status.yaml`
+- `{implementation_artifacts}/*.md` matching pattern `<digit>-<digit>-*.md` (per-story files)
+
+If any exist, log INFO — "Pre-existing artifact found: {filename}. The pipeline is idempotent: `xiaoma-sprint-planning` will smart-detect existing story files and `xiaoma-create-story` skips stories already at `ready-for-dev`."
+
+Also informational scan for optional context (NOT required for this pipeline, but if present they enrich story Dev Notes):
+
+- `{planning_artifacts}/requirements-analysis.md`
+- `{planning_artifacts}/current-architecture-analysis.md`
+- `{planning_artifacts}/architecture.md`
+- `{planning_artifacts}/*ux*.md`
+
+Set pipeline-level flag `{architecture_absent}`:
+
+- If NONE of `{planning_artifacts}/architecture.md`, `{planning_artifacts}/current-architecture-analysis.md`, or any `{planning_artifacts}/*architecture*.md` exists: set `{architecture_absent}` = `true`
+- Otherwise: set `{architecture_absent}` = `false`
+
+If `{architecture_absent}` == `true`: log INFO — "No architecture documents found. Story Dev Notes will be derived solely from prd.md + epics.md. The Data Model subsection of each story will be flagged 'TBD by architect' wherever no schema clues exist in those sources."
+
+This flag is consumed by step-04 — it is passed into every Agent subprocess so each story-creation Agent knows whether to expect architectural context or to default the Data Model subsection to `TBD by architect`.
+
+### 5. Initialize Master Pipeline State Variables
+
+- `{prd_to_stories_status}` = "initialized"
+- `{steps_completed}` = 1
+- `{phase_1_status}` = "pending"
+- `{phase_2_status}` = "pending"
+- `{phase_3_status}` = "pending"
+- `{stories_created}` = 0
+- `{stories_failed}` = 0
+- `{failed_story_keys}` = []
+- `{story_results}` = []
+- `{architecture_absent}` = set by Section 4's scan above (`true` or `false`)
+- Ensure `{planning_artifacts}/` directory exists
+- Ensure `{implementation_artifacts}/` directory exists
+
+### 6. Load Context
+
+- Load `project_context` = `**/project-context.md` (if exists)
+
+### 7. Output Environment Summary
+
+**Output:**
+
+```
+═══════════════════════════════════════════════════
+  Auto PRD-to-Stories Pipeline — INITIALIZED
+═══════════════════════════════════════════════════
+
+Project: {project_name}
+User: {user_name}
+Date: {date}
+Language: {communication_language} / {document_output_language}
+
+📋 Input PRD: {planning_artifacts}/prd.md
+📂 Planning Artifacts: {planning_artifacts}/
+📂 Implementation Artifacts: {implementation_artifacts}/
+
+Pipeline Phases (stops BEFORE development):
+  ⬜ Phase 1 — Create Epics (PM xiaochan, delegates to xiaoma-create-epics-and-stories)
+  ⬜ Phase 2 — Sprint Planning Bridge (SM xiaomin, delegates to xiaoma-sprint-planning)
+  ⬜ Phase 3 — Batch Create Story Files (SM xiaomin, N Agent subprocesses)
+
+Finalization: Step 5 — Unified completion report for business-architect review
+
+⚠️  This pipeline does NOT enter development, code-review, or testing.
+   Output is the input to a manual quality review by business architects.
+
+Master Step: 1/5
+═══════════════════════════════════════════════════
+```
+
+**Auto-Proceed:** YES — Do NOT wait for user input. Proceed to step-02.
+
+---
+
+## NEXT STEP
+
+**NEXT:** Read fully and follow: `{project-root}/_xiaoma/xmc/workflows/5-full-pipeline/auto-prd-to-stories/steps/step-02-create-epics.md`
+
+---
+
+## SUCCESS METRICS
+
+- Config loaded and all variables resolved
+- `prd.md` exists and is non-empty
+- All three delegated sub-workflows accessible
+- Master state variables initialized
+- Environment summary output
+
+## FAILURE MODES
+
+- Missing `prd.md` (cannot proceed without input)
+- Missing XiaoMa config (installation incomplete)
+- Missing sub-workflow files (installation incomplete)
+- Directory creation failures

package/src/xmc/workflows/5-full-pipeline/auto-prd-to-stories/steps/step-02-create-epics.md

@@ -0,0 +1,156 @@
+---
+name: 'step-02-create-epics'
+description: 'Phase 1 — PM role creates epics and stories from prd.md by delegating to create-epics-and-stories workflow'
+nextStepFile: './step-03-bridge-sprint-planning.md'
+---
+
+# Step 2 of 5: Create Epics and Stories
+
+**Goal:** Break down the input `prd.md` into epics and user stories using the existing `xiaoma-create-epics-and-stories` workflow. Produces `{planning_artifacts}/epics.md`.
+
+**Role:** PM (xiaochan) — Product Manager + Requirements Specialist
+
+---
+
+## EXECUTION SEQUENCE
+
+### 1. Announce Phase Entry
+
+**Output:**
+
+```
+═══════════════════════════════════════════════════
+  Phase 1 of 3 — Create Epics from PRD
+═══════════════════════════════════════════════════
+
+Switching role to PM (xiaochan)...
+Delegating to xiaoma-create-epics-and-stories workflow.
+───────────────────────────────────────────────────
+```
+
+### 2. Switch to PM Role
+
+Adopt the PM (xiaochan) persona:
+
+- You are xiaochan, the Product Manager + Requirements Specialist
+- Your purpose is to break down the input PRD into well-structured epics and user stories
+- Communication style: Precise, structured, stakeholder-aware
+- Focus on complete feature coverage and actionable story granularity
+
+### 3. Load Required Context
+
+Load these artifacts in order:
+
+1. **PRIMARY SOURCE** — Read the FULL file from beginning to end
+   - File: `{planning_artifacts}/prd.md` (validated in step-01)
+   - Requirement: Complete read, no skipping
+
+2. **OPTIONAL SECONDARY SOURCES** — Read each fully IF the file exists; if not present, SKIP silently
+   - File: `{planning_artifacts}/requirements-analysis.md` — Optional
+   - File: `{planning_artifacts}/current-architecture-analysis.md` — Optional
+   - File: `{planning_artifacts}/architecture.md` — Optional
+
+   **Note:** Unlike the auto-requirements-pipeline's step-06 which treats these as required, this pipeline treats them as opportunistic enrichment. The new workflow's contract is "PRD-only input is sufficient."
+
+### 4. Verify Workflow Prerequisites
+
+Before delegating, verify the required workflow exists:
+
+1. Check that `{project-root}/_xiaoma/xmc/workflows/3-solutioning/xiaoma-create-epics-and-stories/workflow.md` is readable
+2. If workflow is missing:
+   - HALT — "Required workflow not found: `xiaoma-create-epics-and-stories`. Verify XiaoMa installation is complete."
+
+### 5. Execute Epic and Story Creation (Delegate)
+
+Invoke the `xiaoma-create-epics-and-stories` skill to create the epic breakdown.
+
+**CRITICAL PIPELINE MODE INSTRUCTIONS:**
+
+- This is running in **automated pipeline mode** — do NOT pause for user input
+- The create-epics-and-stories workflow has 4 step files: validate-prerequisites → design-epics → create-stories → final-validation
+- When ANY step presents a menu with **[C] Continue**, automatically select **C** to proceed to the next step
+- Step 1 (validate-prerequisites) will ask the user to confirm discovered documents and extracted requirements — automatically confirm and continue
+- Step 1 asks "Do these extracted requirements accurately represent what needs to be built?" — auto-confirm YES and select **C**
+- When the workflow asks for user decisions on epic boundaries or story granularity, make reasonable decisions based on the PRD content
+- Ensure every PRD feature is covered by at least one epic
+- Each epic should have well-defined user stories with BDD acceptance criteria (Given/When/Then)
+- Output file MUST be: `{planning_artifacts}/epics.md`
+
+### 6. Verify Epic Output
+
+After create-epics-and-stories workflow completes:
+
+1. Verify `{planning_artifacts}/epics.md` exists and is non-empty
+2. Verify all PRD features are covered by at least one epic (sanity check)
+3. Verify each epic has at least one user story
+4. Verify stories have acceptance criteria (look for "Given"/"When"/"Then" or equivalent)
+5. **FR Coverage Map post-condition (CRITICAL):**
+   - Search `epics.md` for a section heading matching `### FR Coverage Map` (or `## FR Coverage Map`)
+   - The section MUST exist and MUST contain at least one non-placeholder mapping line (something like `FR-N → Epic X / Story X.Y`, NOT a raw `{{requirements_coverage_map}}` placeholder)
+   - **Why:** the underlying create-epics-and-stories template includes this section but the LLM occasionally skips filling it in. Without a populated coverage map, step-05's reverse-traceability check loses its anchor and the architect cannot easily verify "no PRD feature was dropped".
+   - If the section is missing OR contains an unfilled `{{...}}` placeholder:
+     - **First retry:** Re-invoke the create-epics-and-stories workflow with an explicit instruction: "The previous run left `### FR Coverage Map` empty or as a placeholder. You MUST extract every FR/feature from prd.md and produce concrete `FR-N → Epic X / Story X.Y` mappings."
+     - If the second run also fails to populate it: do NOT HALT — instead, log WARNING and continue. step-05 will fall back to grepping FR references inside story files and report any uncovered FRs as Structural Warnings.
+6. Also extract a tally for step-05:
+   - `{epics_fr_count}` = number of distinct FR-style references in the FR Coverage Map (e.g., `FR-1`, `FR-2`, … or `PRD §x.y` patterns). Save this for step-05's denominator.
+
+If epic creation failed at the file-level (epics.md missing or unreadable):
+
+- **Retry once:** Re-invoke the workflow with the same instructions
+- If still fails after retry: HALT — "Epic and story creation failed twice. Check `prd.md` for completeness and structural clarity before re-running."
+
+### 7. Output Phase 1 Summary
+
+Count epics and stories from generated `epics.md`:
+
+- `{epic_count}` = number of "## Epic N:" headings
+- `{story_count}` = number of "### Story N.N:" headings
+
+**Output:**
+
+```
+───────────────────────────────────────────────────
+  ✅ Phase 1 Complete — Epics Created
+───────────────────────────────────────────────────
+
+Output: {planning_artifacts}/epics.md
+  Total Epics: {epic_count}
+  Total Stories: {story_count}
+
+Phase 1 Status: success
+Master Step: 2/5
+
+Proceeding to Phase 2 — Sprint Planning Bridge...
+───────────────────────────────────────────────────
+```
+
+### 8. Pipeline Status Update
+
+- `{phase_1_status}` = "success"
+- `{prd_to_stories_status}` = "bridge-phase"
+- `{steps_completed}` = 2
+
+**Auto-Proceed:** YES — Do NOT wait for user input. Immediately proceed to step-03.
+
+---
+
+## NEXT STEP
+
+**NEXT:** Read fully and follow: `{project-root}/_xiaoma/xmc/workflows/5-full-pipeline/auto-prd-to-stories/steps/step-03-bridge-sprint-planning.md`
+
+---
+
+## SUCCESS METRICS
+
+- `{planning_artifacts}/epics.md` created with complete epic breakdown
+- All PRD features covered by at least one epic
+- Each epic has user stories with BDD acceptance criteria
+- Story dependencies identified (where applicable)
+
+## FAILURE MODES
+
+- PRD features not fully covered by epics
+- Stories missing acceptance criteria
+- Epic boundaries too broad or too narrow
+- Workflow not writing output file
+- Two consecutive failures of the delegated workflow