bmm-opencode 1.3.1 → 1.3.3

package/.opencode/agents/party-mode.md

@@ -365,6 +365,7 @@ Provide your expert perspective on this topic. Think deeply and thoroughly.
 - `deep [topic]` - Deep dive on specific topic
 - `vote [options]` - Get agent votes on decision
 - `summarize` - Get executive summary
+- `report` - End meeting and generate summary report file (choose summary or detailed)
 - `exit` - End meeting
 ```
 
@@ -479,6 +480,7 @@ meeting_state:
 | `attendees` | Show current participants | |
 | `summarize` | Executive summary | |
 | `actions` | List action items | |
+| `report` | End meeting + generate report file | `report`, `report summary`, `report detailed` |
 | `exit` / `end` | End meeting | |
 
 ---
@@ -560,10 +562,168 @@ for task_id in active_tasks:
 
 ---
 
+## Meeting Report Generation
+
+When user types `report`, or when exiting via `exit`/`end`:
+
+### Report Trigger Flow
+
+```
+User: report
+
+Response:
+📝 **Generate Meeting Report**
+
+Choose report format:
+- [S] **Summary** - Executive summary with key decisions and action items (~1 page)
+- [D] **Detailed** - Full meeting minutes with all agent contributions, discussion threads, and decision rationale (~3-5 pages)
+
+User: S (or D)
+
+Response:
+📄 Generating report...
+✅ Meeting report saved to: {output_folder}/meeting-reports/meeting-report-{YYYY-MM-DD-HHmm}.md
+```
+
+### Report File Structure
+
+**Summary Report** (`report` or `report summary`):
+
+```markdown
+# Meeting Report - {meeting_topic}
+
+**Date**: {date}
+**Duration**: {duration}
+**Participants**: {list of agent names with icons}
+**Facilitator**: Party Mode
+
+---
+
+## Executive Summary
+[2-3 sentence synthesis of the meeting]
+
+## Key Decisions
+| # | Decision | Rationale | Owner |
+|---|----------|-----------|-------|
+| 1 | [decision] | [why] | [agent] |
+
+## Action Items
+| # | Action | Owner | Priority | Deadline |
+|---|--------|-------|----------|----------|
+| 1 | [task] | [agent] | 🔴/🟡/🟢 | [if discussed] |
+
+## Open Questions
+- [unresolved question 1]
+- [unresolved question 2]
+
+## Next Steps
+- [recommended follow-up 1]
+- [recommended follow-up 2]
+```
+
+**Detailed Report** (`report detailed`):
+
+```markdown
+# Meeting Report - {meeting_topic}
+
+**Date**: {date}
+**Duration**: {duration}
+**Participants**: {list of agent names with icons and titles}
+**Facilitator**: Party Mode
+**Discussion Rounds**: {count}
+
+---
+
+## Executive Summary
+[2-3 sentence synthesis of the meeting]
+
+## Meeting Agenda & Context
+[Original meeting topic and background provided by user]
+
+## Discussion Timeline
+
+### Round 1: [Topic/Question]
+**Participants**: [agents involved]
+
+#### {icon} {Agent Name} ({title})
+[Full contribution from this agent]
+
+#### {icon} {Agent Name} ({title})
+[Full contribution from this agent]
+
+**Round Outcome**: [What was concluded in this round]
+
+---
+
+### Round 2: [Topic/Question]
+[... same structure for each discussion round ...]
+
+---
+
+## Synthesis
+
+### ✅ Consensus Points
+- [What agents agreed on, with supporting agents listed]
+
+### ⚠️ Divergent Views
+| Topic | Perspective A | Perspective B | Resolution |
+|-------|--------------|--------------|------------|
+| [topic] | {Agent}: [view] | {Agent}: [view] | [how resolved or still open] |
+
+### Decision Log
+| # | Decision | Rationale | Proposed By | Supported By |
+|---|----------|-----------|-------------|-------------|
+| 1 | [decision] | [reasoning] | [agent] | [agents] |
+
+## Action Items
+| # | Action | Owner | Priority | Deadline | Context |
+|---|--------|-------|----------|----------|---------|
+| 1 | [task] | [agent] | 🔴/🟡/🟢 | [if discussed] | [brief context] |
+
+## Open Questions & Risks
+| # | Question/Risk | Raised By | Impact | Suggested Resolution |
+|---|--------------|-----------|--------|---------------------|
+| 1 | [question] | [agent] | [impact] | [suggestion] |
+
+## Recommendations for Follow-up
+- [Specific follow-up with recommended agents]
+
+## Appendix: Agent Roster
+| Agent | Title | Key Contribution |
+|-------|-------|-----------------|
+| {icon} {name} | {title} | [one-line summary] |
+```
+
+### Report Generation Rules
+
+1. **On `report` command**: Ask user to choose Summary [S] or Detailed [D]
+2. **On `report summary`**: Generate summary report directly
+3. **On `report detailed`**: Generate detailed report directly
+4. **On `exit`/`end`**: Ask "Would you like to generate a meeting report before ending? [S] Summary / [D] Detailed / [N] No, just exit"
+5. **File output**: Use `write` tool to save the report as a markdown file
+6. **File path**: `{output_folder}/meeting-reports/meeting-report-{YYYY-MM-DD-HHmm}.md` (create directory if needed)
+7. **After report saved**: Show the file path and then proceed with exit if triggered by exit command
+
+---
+
 ## Graceful Exit
 
 When user types `exit`, `end`, or `goodbye`:
 
+**Step 1: Offer report generation**
+
+```
+📝 Before we wrap up — would you like to generate a meeting report?
+
+- [S] **Summary Report** - Key decisions + action items (~1 page)
+- [D] **Detailed Report** - Full meeting minutes with all contributions (~3-5 pages)
+- [N] **No thanks** - Just end the meeting
+```
+
+**Step 2: Generate report if requested** (see Meeting Report Generation section above)
+
+**Step 3: Show meeting conclusion**
+
 ```
 ## 🎊 Meeting Concluded
 
@@ -572,6 +732,7 @@ When user types `exit`, `end`, or `goodbye`:
 - **Duration**: {duration}
 - **Participants**: {list of agents}
 - **Discussion Rounds**: {count}
+- **Report**: {file path if generated, or "Not generated"}
 
 ### Key Decisions Made
 1. [decision 1]
@@ -592,8 +753,7 @@ When user types `exit`, `end`, or `goodbye`:
 
 ---
 
-Thank you for using Party Mode! 
-Meeting notes can be saved with `save [filename]`.
+Thank you for using Party Mode! 🎉
 ```
 
 ---

package/.opencode/skills/bmad-bmm-dev-team-mode/SKILL.md

@@ -127,6 +127,79 @@ Choose:
 | `max_retry_per_story` | 3 | Retries before blocking |
 | `dev_category` | deep | delegate_task category for dev |
 | `review_category` | ultrabrain | delegate_task category for review |
+| `delegated_mode` | false | Auto-detected when pre-filled params present in prompt |
+
+---
+
+## DELEGATED MODE (Sisyphus Ultrawork Support)
+
+When this workflow is launched via `delegate_task` by an orchestrator (e.g., Sisyphus ultrawork), interactive prompts cannot reach the user. Delegated mode solves this.
+
+### Detection
+
+Delegated mode activates automatically when the prompt contains a `PRE-SELECTED:` block:
+
+```
+PRE-SELECTED:
+- epic: epic-2
+- mode: autopilot | continue | select
+- execution: parallel | sequential
+- max_parallel: 3
+```
+
+### Behavior Changes in Delegated Mode
+
+| Step | Normal Mode | Delegated Mode |
+|------|-------------|----------------|
+| Step 1 Selection 1 | Ask user: epic choice | Use `pre-selected.epic` |
+| Step 1 Selection 2 | Ask user: execution mode | Use `pre-selected.execution` |
+| Step 2 Plan confirm | Ask user: proceed? | Auto-proceed (Y) |
+| Step 4 Needs-fixes | Ask user: auto-fix? | Auto-fix always |
+| Step 5 Continue? | Ask user: continue/pause/exit | Auto-continue (autopilot behavior) |
+
+### Pre-Fill Format (for orchestrators)
+
+```
+TASK: Run dev-team-mode
+
+PRE-SELECTED:
+- epic: epic-2
+- mode: autopilot
+- execution: parallel
+- max_parallel: 3
+
+CONTEXT:
+- sprint_status: {path to sprint-status.yaml}
+- story_dir: {path to story files}
+- project_context: {path to project-context.md}
+```
+
+### Structured Output (returned to orchestrator)
+
+When delegated mode completes or exits, return a structured summary:
+
+```yaml
+DEV_TEAM_MODE_RESULT:
+  status: completed | blocked | paused | error
+  epic: epic-2
+  loops: 3
+  stories:
+    done: [1-1-setup, 1-2-auth, 1-3-data]
+    failed: []
+    blocked: [1-4-integration]
+    remaining: [1-5-polish]
+  summary: "3 stories completed, 1 blocked on missing API dependency"
+  action_required: "Resolve 1-4 dependency on external API, then re-run"
+```
+
+### Context Management for Long Loops
+
+In delegated mode, after each loop (Step 5), discard raw agent outputs and retain only:
+- Per-story status (done/failed/blocked + one-line reason)
+- Cumulative stats
+- Current sprint-status.yaml state
+
+This prevents context window overflow during multi-loop autopilot runs.
 
 ---
 

package/README.md

@@ -163,6 +163,7 @@ curl -s https://raw.githubusercontent.com/Jack-R-Hong/BMM-opencode/refs/heads/ma
 
 | Version | Agents | Skills | Changes |
 |---------|--------|--------|---------|
+| 1.3.2 | 19 | 62 | Dev-team-mode: delegated mode for orchestrator support |
 | 1.3.1 | 19 | 62 | Party-mode: agent titles with first impressions |
 | 1.3.0 | 19 | 62 | Added `dev-team-mode` parallel development orchestrator |
 | 1.2.0 | 19 | 61 | Added `party-mode`, `gen-subagent`, upgrade docs |

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "bmm-opencode",
-  "version": "1.3.1",
+  "version": "1.3.3",
   "description": "BMAD-METHOD agents and skills for OpenCode - AI agent framework with 19 specialized agents and 62 workflow skills",
   "main": "dist/index.js",
   "types": "dist/index.d.ts",