bmad-auto-copilot 1.0.6 → 1.1.0

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "bmad-auto-copilot",
-  "version": "1.0.6",
+  "version": "1.1.0",
   "description": "Automated BMAD story development loop using GitHub Copilot CLI — install into any BMAD project",
   "keywords": [
     "bmad",

package/templates/bmad-loop.sh

@@ -178,13 +178,19 @@ get_next_action() {
         echo "[DEBUG] Review: $review_count | Ready: $ready_count | Backlog: $backlog_count | In-Progress: $in_progress_count | Done: $done_count / $total_count" >&2
     fi
 
-    # Priority order: review > ready-for-dev > backlog
+    # Priority order: review > in-progress > ready-for-dev > backlog
     if [[ $review_count -gt 0 ]]; then
         log "[NEXT] Found story in REVIEW state → code-review (Dev agent)" "green" >&2
         echo "code-review"
         return
     fi
 
+    if [[ $in_progress_count -gt 0 ]]; then
+        log "[NEXT] Found story in IN-PROGRESS state → dev-story resume (Dev agent)" "green" >&2
+        echo "dev-story"
+        return
+    fi
+
     if [[ $ready_count -gt 0 ]]; then
         log "[NEXT] Found story in READY-FOR-DEV state → dev-story (Dev agent)" "green" >&2
         echo "dev-story"
@@ -203,11 +209,6 @@ get_next_action() {
         return
     fi
 
-    # Check for in-progress (waiting state)
-    if [[ $in_progress_count -gt 0 ]]; then
-        log "[WAIT] Found $in_progress_count story(ies) in-progress" "yellow" >&2
-    fi
-
     echo "wait"
 }
 

package/templates/prompts/code-review.md

@@ -1,32 +1,44 @@
 You are an automated BMAD workflow executor. Your task is to execute the code-review workflow.
 
-IMPORTANT: This is a non-interactive automated execution. Do NOT show menus, greetings, or ask questions. Execute the workflow immediately.
+IMPORTANT: This is a non-interactive automated execution. Do NOT show menus, greetings, or ask questions. Execute the workflow immediately. Do NOT pause between steps or ask "Continue to next step?" — execute ALL steps continuously.
 
 Timestamp: {TIMESTAMP}
 
-## Step 1: Load Workflow
+## Step 1: Load Configuration and Workflow
+Read the BMAD config from: `_bmad/bmm/config.yaml` — resolve all variables ({project-root}, {output_folder}, etc.)
 Read the workflow definition from: `_bmad/bmm/workflows/4-implementation/code-review/workflow.yaml`
 Read the workflow instructions from: `_bmad/bmm/workflows/4-implementation/code-review/instructions.xml`
 Read the workflow checklist from: `_bmad/bmm/workflows/4-implementation/code-review/checklist.md`
 
 ## Step 2: Find Next Story
-Read `_bmad-output/implementation-artifacts/sprint-status.yaml` and find the FIRST story in `review` state (scan top to bottom).
+Read `_bmad-output/implementation-artifacts/sprint-status.yaml` and find the FIRST story in `review` state (scan top to bottom, exclude retrospectives).
 
 ## Step 3: Load Story File
 Read the corresponding story file from `_bmad-output/implementation-artifacts/stories/`.
+Parse sections: Story, Acceptance Criteria, Tasks/Subtasks, Dev Agent Record (File List, Change Log).
 
-## Step 4: Adversarial Code Review
+## Step 4: Adversarial Code Review with Auto-Fix
 Perform a thorough ADVERSARIAL code review following the workflow instructions:
-- Find 3-10 specific problems across: code quality, test coverage, architecture compliance, security, performance
-- Auto-fix ALL fixable issues
+- Use `git status --porcelain`, `git diff`, `git diff --cached` to discover actual changes
+- Cross-reference story File List with git reality — note discrepancies
+- Verify EVERY task marked [x] is actually implemented
+- Verify EVERY Acceptance Criterion is actually satisfied
+- Find 3-10 specific problems: code quality, test coverage, architecture compliance, security, performance
+- **AUTOMATICALLY FIX all HIGH and MEDIUM issues** (this is option [1] from instructions.xml Step 4 — always choose this)
 - Run all tests to verify fixes pass
+- Update the story File List and Dev Agent Record with any fixes applied
 - NEVER accept "looks good" without finding real issues
 
 ## Step 5: Update Sprint Status
-Update `_bmad-output/implementation-artifacts/sprint-status.yaml` — change the story status from `review` to `done`.
+Based on review outcome:
+- If ALL HIGH and MEDIUM issues are fixed AND all ACs implemented: update status from `review` to `done`
+- If issues remain unfixable: update status from `review` to `in-progress` (will re-enter dev cycle)
+- Preserve ALL comments and structure in `_bmad-output/implementation-artifacts/sprint-status.yaml`
 
 ## Rules
 - Use best judgment for ALL decisions — never ask the user
-- Document issues you cannot auto-fix
+- Execute ALL steps without pausing — this is YOLO mode (no confirmation between steps)
+- When instructions.xml has `<ask>` tags, always choose "Fix them automatically" (option 1)
+- Document issues you cannot auto-fix in the story file
 - All tests must pass after any fixes
 - Follow the workflow checklist to verify completeness

package/templates/prompts/create-story.md

@@ -1,29 +1,36 @@
 You are an automated BMAD workflow executor. Your task is to execute the create-story workflow.
 
-IMPORTANT: This is a non-interactive automated execution. Do NOT show menus, greetings, or ask questions. Execute the workflow immediately.
+IMPORTANT: This is a non-interactive automated execution. Do NOT show menus, greetings, or ask questions. Execute the workflow immediately. Do NOT pause between steps or ask "Continue to next step?" — execute ALL steps continuously.
 
 Timestamp: {TIMESTAMP}
 
-## Step 1: Load Workflow
+## Step 1: Load Configuration and Workflow
+Read the BMAD config from: `_bmad/bmm/config.yaml` — resolve all variables ({project-root}, {output_folder}, etc.)
 Read the workflow definition from: `_bmad/bmm/workflows/4-implementation/create-story/workflow.yaml`
 Read the workflow instructions from: `_bmad/bmm/workflows/4-implementation/create-story/instructions.xml`
 Read the workflow template from: `_bmad/bmm/workflows/4-implementation/create-story/template.md`
 Read the workflow checklist from: `_bmad/bmm/workflows/4-implementation/create-story/checklist.md`
 
 ## Step 2: Find Next Story
-Read `_bmad-output/implementation-artifacts/sprint-status.yaml` and find the FIRST story in `backlog` state (scan top to bottom).
+Read `_bmad-output/implementation-artifacts/sprint-status.yaml` and find the FIRST story in `backlog` state (scan top to bottom, exclude retrospectives).
 
-## Step 3: Load Epic Context
-Read the corresponding epic file from `_bmad-output/planning-artifacts/` to understand story requirements, acceptance criteria, and BDD scenarios.
+## Step 3: Load and Analyze Context (Exhaustive)
+Follow the instructions.xml discover_inputs protocol:
+- Read the corresponding epic file from `_bmad-output/planning-artifacts/`
+- Read PRD, architecture docs, and any referenced project context artifacts
+- Analyze previous story files for context continuity (patterns, conventions, decisions)
+- Understand story requirements, acceptance criteria, and BDD scenarios
 
 ## Step 4: Create Story File
-Follow the workflow instructions and template to create the story file in `_bmad-output/implementation-artifacts/stories/`. The story file must include all acceptance criteria, tasks, subtasks, and implementation notes.
+Follow the workflow instructions and template to create the story file in `_bmad-output/implementation-artifacts/stories/`. The story file must include all acceptance criteria, tasks, subtasks, dev notes, and implementation guidance.
 
 ## Step 5: Update Sprint Status
-Update `_bmad-output/implementation-artifacts/sprint-status.yaml` — change the story status from `backlog` to `ready-for-dev`.
+Update `_bmad-output/implementation-artifacts/sprint-status.yaml` — change the story status from `backlog` to `ready-for-dev`. Preserve ALL comments and existing structure in the YAML file.
 
 ## Rules
 - Use best judgment for ALL decisions — never ask the user
 - Do NOT implement the story code — only create the story file
 - Follow the workflow template precisely
+- Execute ALL steps without pausing — this is YOLO mode (no confirmation between steps)
+- When instructions.xml has `<ask>` tags, use best judgment instead of asking
 - If blocking issues arise, document them in the story file and continue

package/templates/prompts/dev-story.md

@@ -1,33 +1,51 @@
 You are an automated BMAD workflow executor. Your task is to execute the dev-story workflow.
 
-IMPORTANT: This is a non-interactive automated execution. Do NOT show menus, greetings, or ask questions. Execute the workflow immediately.
+IMPORTANT: This is a non-interactive automated execution. Do NOT show menus, greetings, or ask questions. Execute the workflow immediately. Do NOT pause between steps or ask "Continue to next step?" — execute ALL steps continuously until the story is COMPLETE.
 
 Timestamp: {TIMESTAMP}
 
-## Step 1: Load Workflow
+## Step 1: Load Configuration and Workflow
+Read the BMAD config from: `_bmad/bmm/config.yaml` — resolve all variables ({project-root}, {output_folder}, etc.)
 Read the workflow definition from: `_bmad/bmm/workflows/4-implementation/dev-story/workflow.yaml`
 Read the workflow instructions from: `_bmad/bmm/workflows/4-implementation/dev-story/instructions.xml`
 Read the workflow template (if exists) from: `_bmad/bmm/workflows/4-implementation/dev-story/`
 Read the workflow checklist from: `_bmad/bmm/workflows/4-implementation/dev-story/checklist.md`
 
 ## Step 2: Find Next Story
-Read `_bmad-output/implementation-artifacts/sprint-status.yaml` and find the FIRST story in `ready-for-dev` state (scan top to bottom).
+Read `_bmad-output/implementation-artifacts/sprint-status.yaml` and find the FIRST story in `ready-for-dev` or `in-progress` state (scan top to bottom, exclude retrospectives).
+- If `in-progress` found: this is a resumed session — continue where the previous session left off.
+- If `ready-for-dev` found: this is a fresh implementation.
 
 ## Step 3: Load Story File
 Read the corresponding story file from `_bmad-output/implementation-artifacts/stories/`.
+- Check for "Senior Developer Review (AI)" section — if present, this is a review continuation. Extract pending review items and address them first.
 
-## Step 4: Implement
+## Step 4: Update Sprint Status to In-Progress
+Update `_bmad-output/implementation-artifacts/sprint-status.yaml` — change the story status from `ready-for-dev` to `in-progress` (if not already in-progress). Preserve ALL comments and structure.
+
+## Step 5: Implement
 Follow the workflow instructions to implement ALL tasks and subtasks defined in the story:
-- Write production code for every task
+- Follow the red-green-refactor TDD cycle as specified in instructions.xml
+- Write failing tests first, then minimal code to pass, then refactor
 - Write comprehensive unit tests — tests must pass 100%
 - Follow ALL acceptance criteria precisely
-- Update the story file to track task completion
+- Update the story file to track task completion (mark [x] only when verified)
+- Update the File List section with all new/modified/deleted files
+- Execute continuously — do NOT stop at milestones or for progress updates
 
-## Step 5: Update Sprint Status
-Update `_bmad-output/implementation-artifacts/sprint-status.yaml` — change the story status from `ready-for-dev` to `review`.
+## Step 6: Update Sprint Status to Review
+After ALL tasks are complete and ALL tests pass:
+- Run the full test suite to ensure no regressions
+- Validate definition-of-done checklist
+- Update `_bmad-output/implementation-artifacts/sprint-status.yaml` — change the story status from `in-progress` to `review`
+- Preserve ALL comments and structure in the YAML file
 
 ## Rules
 - Use best judgment for ALL decisions — never ask the user
+- Execute ALL steps without pausing — this is YOLO mode (no confirmation between steps)
+- When instructions.xml has `<ask>` tags, use best judgment instead of asking
 - Follow the workflow checklist to verify completeness
 - Ensure all existing tests continue to pass
+- NEVER mark a task [x] unless it is truly complete with passing tests
+- Do NOT stop because of milestones or significant progress — continue until ALL tasks are done
 - If blocking issues arise, document them and continue with what you can