agileflow 3.0.1 → 3.1.0

package/src/core/commands/legal/audit.md

@@ -0,0 +1,446 @@
+---
+description: Multi-agent legal risk analysis with consensus voting for finding compliance gaps
+argument-hint: "[file|directory] [DEPTH=quick|deep] [FOCUS=privacy|terms|a11y|licensing|consumer|security|ai|content|international|all]"
+compact_context:
+  priority: high
+  preserve_rules:
+    - "ACTIVE COMMAND: /agileflow:legal:audit - Multi-agent legal risk analysis"
+    - "CRITICAL: Deploy analyzers IN PARALLEL in ONE message with multiple Task calls"
+    - "CRITICAL: Wait for all results before running consensus (use TaskOutput with block=true)"
+    - "CRITICAL: Confidence scoring: CONFIRMED (2+ agree), LIKELY (1 with evidence), INVESTIGATE (1 weak)"
+    - "MUST parse arguments: TARGET (file/dir), DEPTH (quick/deep), FOCUS (privacy|terms|a11y|licensing|consumer|security|ai|content|international|all)"
+    - "Pass consensus all analyzer outputs, let it synthesize the final report"
+  state_fields:
+    - target_path
+    - depth
+    - focus_areas
+    - analyzers_deployed
+    - findings_collected
+---
+
+# /agileflow:legal:audit
+
+Deploy multiple specialized legal risk analyzers in parallel to find compliance gaps, then synthesize results through consensus voting into a prioritized Legal Risk Report.
+
+---
+
+## Quick Reference
+
+```
+/agileflow:legal:audit app/                                # Analyze app directory (quick, core 5 analyzers)
+/agileflow:legal:audit . DEPTH=deep                        # Deep analysis - all 9 analyzers
+/agileflow:legal:audit src/ FOCUS=privacy,a11y             # Focus on specific areas
+/agileflow:legal:audit . DEPTH=deep FOCUS=all              # Comprehensive full audit
+/agileflow:legal:audit app/page.tsx FOCUS=ai               # Check single file for AI compliance
+```
+
+---
+
+## How It Works
+
+```
+┌─────────────────────────────────────────────────────────────┐
+│                 /agileflow:legal:audit                       │
+│                                                              │
+│  1. Parse arguments (target, depth, focus)                   │
+│  2. Deploy analyzers IN PARALLEL                             │
+│  3. Collect all findings                                     │
+│  4. Run consensus coordinator to validate & prioritize       │
+│  5. Generate actionable Legal Risk Report                    │
+└─────────────────────────────────────────────────────────────┘
+
+   ┌────────┐ ┌────────┐ ┌────────┐ ┌────────┐ ┌────────┐
+   │Privacy │ │ Terms  │ │ A11y   │ │License │ │Consumer│
+   └───┬────┘ └───┬────┘ └───┬────┘ └───┬────┘ └───┬────┘
+       │          │          │          │          │
+   ┌───┴────┐ ┌───┴────┐ ┌───┴────┐ ┌───┴────┐
+   │Security│ │  AI    │ │Content │ │ Intl   │  (deep only)
+   └───┬────┘ └───┬────┘ └───┬────┘ └───┬────┘
+       │          │          │          │
+       └──────────┴──────────┼──────────┘
+                             ▼
+               ┌─────────────────────────┐
+               │   Consensus Coordinator │
+               │   (validates, votes,    │
+               │    generates report)    │
+               └─────────────────────────┘
+```
+
+---
+
+## Arguments
+
+| Argument | Values | Default | Description |
+|----------|--------|---------|-------------|
+| TARGET | file/directory | `.` | What to analyze |
+| DEPTH | quick, deep | quick | quick = core 5 analyzers, deep = all 9 |
+| FOCUS | privacy,terms,a11y,licensing,consumer,security,ai,content,international,all | all | Which analyzers to deploy |
+
+---
+
+## Step-by-Step Process
+
+### STEP 1: Parse Arguments
+
+```
+TARGET = first argument or current directory
+DEPTH = quick (default) or deep
+FOCUS = all (default) or comma-separated list
+```
+
+**Analyzer Selection**:
+
+| Condition | Analyzers Deployed |
+|-----------|-------------------|
+| `DEPTH=quick` + `FOCUS=all` | privacy, terms, a11y, licensing, consumer (core 5) |
+| `DEPTH=deep` + `FOCUS=all` | All 9 analyzers |
+| `FOCUS=privacy` | legal-analyzer-privacy only |
+| `FOCUS=terms` | legal-analyzer-terms only |
+| `FOCUS=a11y` | legal-analyzer-a11y only |
+| `FOCUS=licensing` | legal-analyzer-licensing only |
+| `FOCUS=consumer` | legal-analyzer-consumer only |
+| `FOCUS=security` | legal-analyzer-security only |
+| `FOCUS=ai` | legal-analyzer-ai only |
+| `FOCUS=content` | legal-analyzer-content only |
+| `FOCUS=international` | legal-analyzer-international only |
+| `FOCUS=privacy,a11y` | Comma-separated: deploy specified analyzers |
+
+**DEPTH behavior**:
+- `quick` (default): Deploy core 5 analyzers. Focus on CRITICAL/HIGH issues only.
+- `deep`: Deploy all 9 analyzers. Include MEDIUM/LOW findings.
+
+### STEP 2: Deploy Analyzers in Parallel
+
+**CRITICAL**: Deploy ALL selected analyzers in a SINGLE message with multiple Task calls.
+
+**Prompt template for each analyzer**:
+
+```
+TASK: Analyze the following code for {LEGAL_DOMAIN} compliance risks.
+
+TARGET: {file_path or directory}
+DEPTH: {quick|deep}
+
+{For quick depth}: Focus on CRITICAL and HIGH risk issues only. Skip advisory/best-practice items.
+{For deep depth}: Be comprehensive. Include MEDIUM and LOW risk findings.
+
+Read the target files and apply your analysis methodology.
+
+OUTPUT your findings in your standard format (FINDING-N with location, risk level, legal basis, code, explanation, remediation).
+
+If no issues found, output: "No {LEGAL_DOMAIN} compliance issues found in {TARGET}"
+```
+
+**Example deployment (DEPTH=quick, FOCUS=all - deploys core 5)**:
+
+```xml
+<invoke name="Task">
+<parameter name="description">Privacy compliance analysis</parameter>
+<parameter name="prompt">TASK: Analyze the following code for PRIVACY & DATA PROTECTION compliance risks.
+TARGET: src/
+DEPTH: quick
+Focus on CRITICAL and HIGH risk issues only...
+...</parameter>
+<parameter name="subagent_type">legal-analyzer-privacy</parameter>
+<parameter name="run_in_background">true</parameter>
+</invoke>
+
+<invoke name="Task">
+<parameter name="description">Terms & legal docs analysis</parameter>
+<parameter name="prompt">TASK: Analyze the following code for TERMS & LEGAL DOCUMENT compliance risks.
+TARGET: src/
+DEPTH: quick
+...</parameter>
+<parameter name="subagent_type">legal-analyzer-terms</parameter>
+<parameter name="run_in_background">true</parameter>
+</invoke>
+
+<invoke name="Task">
+<parameter name="description">Accessibility compliance analysis</parameter>
+<parameter name="prompt">TASK: Analyze the following code for ACCESSIBILITY compliance risks.
+TARGET: src/
+DEPTH: quick
+...</parameter>
+<parameter name="subagent_type">legal-analyzer-a11y</parameter>
+<parameter name="run_in_background">true</parameter>
+</invoke>
+
+<invoke name="Task">
+<parameter name="description">License compliance analysis</parameter>
+<parameter name="prompt">TASK: Analyze the following code for LICENSING & IP compliance risks.
+TARGET: src/
+DEPTH: quick
+...</parameter>
+<parameter name="subagent_type">legal-analyzer-licensing</parameter>
+<parameter name="run_in_background">true</parameter>
+</invoke>
+
+<invoke name="Task">
+<parameter name="description">Consumer protection analysis</parameter>
+<parameter name="prompt">TASK: Analyze the following code for CONSUMER PROTECTION compliance risks.
+TARGET: src/
+DEPTH: quick
+...</parameter>
+<parameter name="subagent_type">legal-analyzer-consumer</parameter>
+<parameter name="run_in_background">true</parameter>
+</invoke>
+```
+
+**For DEPTH=deep, also deploy**:
+
+```xml
+<invoke name="Task">
+<parameter name="description">Security legal obligations analysis</parameter>
+<parameter name="prompt">TASK: Analyze the following code for SECURITY LEGAL OBLIGATION compliance risks...
+...</parameter>
+<parameter name="subagent_type">legal-analyzer-security</parameter>
+<parameter name="run_in_background">true</parameter>
+</invoke>
+
+<invoke name="Task">
+<parameter name="description">AI compliance analysis</parameter>
+<parameter name="prompt">TASK: Analyze the following code for AI & ALGORITHMIC compliance risks...
+...</parameter>
+<parameter name="subagent_type">legal-analyzer-ai</parameter>
+<parameter name="run_in_background">true</parameter>
+</invoke>
+
+<invoke name="Task">
+<parameter name="description">Content & IP obligations analysis</parameter>
+<parameter name="prompt">TASK: Analyze the following code for CONTENT MODERATION & IP OBLIGATION compliance risks...
+...</parameter>
+<parameter name="subagent_type">legal-analyzer-content</parameter>
+<parameter name="run_in_background">true</parameter>
+</invoke>
+
+<invoke name="Task">
+<parameter name="description">International compliance analysis</parameter>
+<parameter name="prompt">TASK: Analyze the following code for INTERNATIONAL COMPLIANCE risks...
+...</parameter>
+<parameter name="subagent_type">legal-analyzer-international</parameter>
+<parameter name="run_in_background">true</parameter>
+</invoke>
+```
+
+### STEP 3: Collect Results
+
+Wait for all analyzers to complete:
+
+```xml
+<invoke name="TaskOutput">
+<parameter name="task_id">{privacy_id}</parameter>
+<parameter name="block">true</parameter>
+</invoke>
+
+<invoke name="TaskOutput">
+<parameter name="task_id">{terms_id}</parameter>
+<parameter name="block">true</parameter>
+</invoke>
+
+<!-- ... collect all results ... -->
+```
+
+### STEP 4: Run Consensus Coordinator
+
+Pass all analyzer outputs to the consensus coordinator:
+
+```xml
+<invoke name="Task">
+<parameter name="description">Legal audit consensus</parameter>
+<parameter name="prompt">You are the Legal Consensus Coordinator.
+
+TARGET: {target_path}
+DEPTH: {depth}
+
+## Analyzer Outputs
+
+### Privacy Analyzer Results:
+{privacy_output}
+
+### Terms Analyzer Results:
+{terms_output}
+
+### Accessibility Analyzer Results:
+{a11y_output}
+
+### Licensing Analyzer Results:
+{licensing_output}
+
+### Consumer Protection Analyzer Results:
+{consumer_output}
+
+{If deep depth, also include:}
+### Security Legal Obligations Results:
+{security_output}
+
+### AI Compliance Results:
+{ai_output}
+
+### Content & IP Obligations Results:
+{content_output}
+
+### International Compliance Results:
+{international_output}
+
+---
+
+Follow your consensus process:
+1. Detect project type from the codebase
+2. Parse all findings into normalized structure
+3. Group related findings by location
+4. Vote on confidence (CONFIRMED if 2+ agree, LIKELY if 1 with evidence)
+5. Filter by project type relevance
+6. Generate the final Legal Risk Report
+7. Save report to docs/08-project/legal-audits/legal-audit-{YYYYMMDD}.md
+</parameter>
+<parameter name="subagent_type">legal-consensus</parameter>
+</invoke>
+```
+
+### STEP 5: Present Results
+
+After consensus completes, show the report summary and offer next steps:
+
+```xml
+<invoke name="AskUserQuestion">
+<parameter name="questions">[{
+  "question": "Legal audit complete: [N] findings ([critical] Critical, [high] High). [files_count] files analyzed. Project type: [type].",
+  "header": "Next steps",
+  "multiSelect": false,
+  "options": [
+    {"label": "Fix [critical] Critical issues now (Recommended)", "description": "[top_issue_summary] - [legal_basis]"},
+    {"label": "Create stories for all findings", "description": "Track [critical] critical + [high] high priority items in backlog"},
+    {"label": "Re-run with DEPTH=deep on [target]", "description": "Current was quick (5 analyzers) - deep adds Security, AI, Content, International"},
+    {"label": "Save report and done", "description": "Report saved to docs/08-project/legal-audits/"}
+  ]
+}]</parameter>
+</invoke>
+```
+
+---
+
+## Example Output
+
+```
+⚖️  Legal Audit: app/
+══════════════════════════════════════════════════════════════
+
+Deploying 5 legal analyzers (quick mode)...
+✓ Privacy Analyzer
+✓ Terms Analyzer
+✓ Accessibility Analyzer
+✓ Licensing Analyzer
+✓ Consumer Protection Analyzer
+
+Running consensus...
+✓ Consensus complete
+✓ Project type detected: SaaS Application
+
+━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+📊 RISK SUMMARY
+━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+
+| Risk Level | Count |
+|------------|-------|
+| Critical   | 2     |
+| High       | 3     |
+| Medium     | 4     |
+| Low        | 1     |
+
+Total: 10 findings (3 false positives excluded)
+
+━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+🚨 FIX BEFORE LAUNCH
+━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+
+1. No privacy policy but collecting email [CONFIRMED by Privacy, Terms]
+   Location: app/page.tsx:42
+   Legal Basis: GDPR Article 13, CCPA
+   Fix: Add /privacy page and link from footer
+
+2. Images without alt text (12 instances) [CONFIRMED by A11y]
+   Location: components/*.tsx
+   Legal Basis: ADA Title III, WCAG 2.1 AA 1.1.1
+   Fix: Add descriptive alt attributes to all images
+
+━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+⚠️  FIX THIS SPRINT
+━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+
+3. Missing Terms of Service [LIKELY - Terms]
+4. Cookie consent banner absent [LIKELY - Privacy]
+5. GPL dependency in MIT project [LIKELY - Licensing]
+
+[Full report saved to docs/08-project/legal-audits/legal-audit-20260214.md]
+```
+
+---
+
+<!-- COMPACT_SUMMARY_START -->
+## Compact Summary
+
+**Command**: `/agileflow:legal:audit` - Multi-agent legal risk analysis with consensus
+
+**Quick Usage**:
+```
+/agileflow:legal:audit app/                        # Quick scan (core 5 analyzers)
+/agileflow:legal:audit . DEPTH=deep                # All 9 analyzers
+/agileflow:legal:audit src/ FOCUS=privacy,a11y     # Specific areas
+```
+
+**What It Does**: Deploy legal analyzers in parallel → Each finds different compliance gaps → Consensus coordinator validates, filters by project type, prioritizes → Actionable Legal Risk Report
+
+**Analyzers (Core 5 - quick mode)**:
+- `legal-analyzer-privacy` - GDPR, CCPA, cookies, data collection
+- `legal-analyzer-terms` - ToS, disclaimers, refund policies
+- `legal-analyzer-a11y` - ADA, WCAG, Section 508
+- `legal-analyzer-licensing` - Open source licenses, IP, attribution
+- `legal-analyzer-consumer` - Dark patterns, FTC, COPPA
+
+**Analyzers (Deep mode adds 4 more)**:
+- `legal-analyzer-security` - Breach notification, PCI-DSS, encryption obligations
+- `legal-analyzer-ai` - EU AI Act, algorithmic bias, AI disclosure
+- `legal-analyzer-content` - DMCA, Digital Services Act, UGC moderation
+- `legal-analyzer-international` - LGPD, PIPL, data localization, cross-border
+
+**Risk Levels**:
+- CRITICAL: Active lawsuit risk → Fix before launch
+- HIGH: Regulatory fine risk → Fix this sprint
+- MEDIUM: Best practice gap → Backlog
+- LOW: Advisory improvement
+
+**Confidence Levels**:
+- CONFIRMED: 2+ analyzers agree → High priority
+- LIKELY: 1 analyzer with evidence → Medium priority
+- INVESTIGATE: 1 analyzer, weak evidence → Low priority
+
+**Output**: `docs/08-project/legal-audits/legal-audit-{YYYYMMDD}.md`
+<!-- COMPACT_SUMMARY_END -->
+
+---
+
+## Integration with Babysit
+
+When `/agileflow:babysit` completes implementation, it can run a quick legal audit:
+
+```
+📍 Implementation complete. Running quick legal audit...
+
+⚖️  Legal Audit Results:
+━━━━━━━━━━━━━━━━━━━━━━
+✅ No critical legal risks found
+⚠️  1 HIGH issue detected:
+   - app/page.tsx:42 - Collecting email without privacy policy link
+     Confidence: CONFIRMED (Privacy + Terms analyzers)
+
+Fix before launch? [Y/n]
+```
+
+---
+
+## Related Commands
+
+- `/agileflow:logic:audit` - Logic bug analysis (similar architecture)
+- `/agileflow:review` - Code review (includes some compliance checks)
+- `/agileflow:multi-expert` - General multi-expert analysis
+- `/agileflow:verify` - Run tests

package/src/core/commands/logic/audit.md

@@ -239,14 +239,14 @@ After consensus completes, show the report summary and offer next steps:
 ```xml
 <invoke name="AskUserQuestion">
 <parameter name="questions">[{
-  "question": "What would you like to do with the audit results?",
+  "question": "Logic audit: [N] findings ([critical] CONFIRMED, [likely] LIKELY). [files_count] files analyzed.",
   "header": "Next steps",
   "multiSelect": false,
   "options": [
-    {"label": "Fix critical issues now (Recommended)", "description": "Address P0/P1 issues immediately"},
-    {"label": "Create stories for findings", "description": "Add to backlog as user stories"},
-    {"label": "Run deeper analysis", "description": "Re-run with DEPTH=deep for more coverage"},
-    {"label": "Save report and done", "description": "Keep the report, no further action"}
+    {"label": "Fix [critical] CONFIRMED issues now (Recommended)", "description": "[top_issue_summary] at [file:line] - [fix_hint]"},
+    {"label": "Create [N] stories for all findings", "description": "Track [critical] critical + [high] high priority items in backlog"},
+    {"label": "Re-run with DEPTH=deep on [target]", "description": "Current was quick - deep adds P2 issues and comprehensive coverage"},
+    {"label": "Save report and done", "description": "Report saved to docs/08-project/logic-audits/"}
   ]
 }]</parameter>
 </invoke>

package/src/core/commands/review.md

@@ -641,7 +641,13 @@ Analyzing 127 lines changed across 3 files...
 Summary: 1 critical, 3 warnings, 3 good
 ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
 
-Apply auto-fixes for non-critical issues? [Y/n]
+AskUserQuestion:
+  question: "Review: 85/100 score. 1 critical (SQL injection at :45), 3 warnings. How to proceed?"
+  options:
+    - "Fix SQL injection at line 45 first (Recommended)" → "Critical - blocks merge"
+    - "Auto-fix 3 style warnings" → "Safe formatting fixes only"
+    - "View full diff of all fixes" → "Preview 4 changes before applying"
+    - "Skip fixes, save report only" → "Report saved to docs/08-project/code-reviews/"
 ```
 
 ### Success - Clean Review

package/src/core/commands/rpi.md

@@ -392,7 +392,7 @@ Reading plan: `.claude/plans/[slug].md`
 Running tests to verify step 1...
 ```
 
-4. **Complete implementation**:
+4. **Complete implementation** with smart AskUserQuestion:
 ```markdown
 ---
 
@@ -404,13 +404,23 @@ All [N] steps executed successfully.
 - [X] files modified
 - [Y] tests added/updated
 - [Z] tests passing
+```
 
-### Next Actions
-- [ ] Run full test suite: `npm test`
-- [ ] Create PR: `/agileflow:pr-template`
-- [ ] Update story status: `/agileflow:status STORY=US-#### STATUS=in-review`
-
----
+Then present smart next steps:
+```xml
+<invoke name="AskUserQuestion">
+<parameter name="questions">[{
+  "question": "Implementation complete: [N] steps done, [X] files modified, [Z] tests passing. What next?",
+  "header": "Next step",
+  "multiSelect": false,
+  "options": [
+    {"label": "Run full test suite (Recommended)", "description": "Verify all [Z] tests still pass after [X] file changes"},
+    {"label": "🔍 Run logic audit on [X] modified files", "description": "5 analyzers check for edge cases, race conditions, type bugs"},
+    {"label": "Create PR with /agileflow:pr", "description": "[X] files changed across [N] implementation steps"},
+    {"label": "Pause here", "description": "Implementation saved, tests can be run later"}
+  ]
+}]</parameter>
+</invoke>
 ```
 
 ### Step 4: Monitor Context Health
@@ -423,18 +433,22 @@ Throughout execution, monitor for signs of context degradation:
 - Heavy tool output (JSON, logs)
 - Repetitive or generic responses
 
-**If detected, warn**:
-```markdown
-⚠️ **Context Health Warning**
+**If detected, use smart AskUserQuestion**:
 
-This conversation has grown significantly. We may be approaching the "dumb zone" where agent quality degrades.
-
-**Current phase**: [phase]
-**Recommendation**: [Complete phase and start fresh / Save progress and compact]
-
-Would you like to:
-1. Complete current phase and save artifact now
-2. Continue (understanding quality may degrade)
+```xml
+<invoke name="AskUserQuestion">
+<parameter name="questions">[{
+  "question": "⚠️ Context at ~[utilization]% ([phase] phase). [files_read] files read, [tool_calls] tool calls. Save progress?",
+  "header": "Context health",
+  "multiSelect": false,
+  "options": [
+    {"label": "Save [phase] artifact and start fresh (Recommended)", "description": "New conversation starts in smart zone (<40%). Artifact preserves all understanding."},
+    {"label": "Compact context and continue", "description": "Reduce utilization but may lose some nuance"},
+    {"label": "Continue without saving", "description": "⚠️ Quality may degrade - currently at [utilization]%"},
+    {"label": "Pause here", "description": "Progress saved to [artifact_path]"}
+  ]
+}]</parameter>
+</invoke>
 ```
 
 ---
@@ -525,7 +539,13 @@ Research Summary:
 - Patterns: Use existing middleware structure
 - Risks: Token storage, refresh logic
 
-Ready for planning phase? [Y/n]
+AskUserQuestion:
+  question: "Research complete: 4 findings, 2 risks identified. Ready for planning phase?"
+  options:
+    - "Start planning in new conversation (Recommended)" → "Fresh context with research artifact compressed"
+    - "Continue planning in this conversation" → "Context at ~35% - still in smart zone"
+    - "Refine research first" → "4 findings may need deeper investigation"
+    - "Pause here" → "Research saved to docs/10-research/"
 
 ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
 Phase 2: PLAN
@@ -542,7 +562,13 @@ Files to modify: 4
 New files: 2
 Estimated complexity: Medium
 
-Approve plan and proceed to implement? [Y/n]
+AskUserQuestion:
+  question: "Plan ready: 5 steps, 6 files, ~medium complexity. Approve and proceed?"
+  options:
+    - "Approve - start implementing in new conversation (Recommended)" → "Fresh context with plan artifact"
+    - "Approve - implement in this conversation" → "Context at ~50% - proceed with caution"
+    - "Revise step 3 (update login endpoint)" → "Adjust plan before implementing"
+    - "Pause here" → "Plan saved to .claude/plans/"
 
 ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
 Phase 3: IMPLEMENT
@@ -555,6 +581,14 @@ Implementing...
 ✓ Added 12 tests (all passing)
 
 🎉 Implementation complete!
+
+AskUserQuestion:
+  question: "Implementation complete: 5 steps, 4 files modified, 12 tests passing. What next?"
+  options:
+    - "Run full test suite (Recommended)" → "Verify all 12 tests pass"
+    - "🔍 Run logic audit on 4 modified files" → "5 analyzers check for edge cases"
+    - "Create PR" → "4 files changed across 5 steps"
+    - "Pause here" → "Implementation saved"
 ```
 
 ### Success - Phase Transition
@@ -567,12 +601,13 @@ Current: RESEARCH → Next: PLAN
 
 ⚠️ Context Health: 65% (approaching threshold)
 
-Before transitioning:
-1. Document key findings
-2. Save research notes
-3. Plan is LEVERAGE - quality > speed now
-
-Proceed with transition? [Y/n]
+AskUserQuestion:
+  question: "Context at 65% - approaching dumb zone. Save artifact and transition?"
+  options:
+    - "Save artifact and start new conversation (Recommended)" → "Fresh context in smart zone (<40%)"
+    - "Continue in this conversation" → "⚠️ Quality may degrade past 70%"
+    - "Compact context first" → "Reduce utilization, then continue"
+    - "Pause here" → "Progress saved, resume later"
 ```
 
 ### Error - Wrong Phase

package/src/core/commands/sprint.md

@@ -95,17 +95,18 @@ TaskCreate/TaskUpdate:
 </invoke>
 ```
 
-AskUserQuestion:
+AskUserQuestion (use actual counts and data from analysis):
 ```xml
 <invoke name="AskUserQuestion">
 <parameter name="questions">[{
-  "question": "Commit this sprint plan: 5 stories, 8.5 days?",
-  "header": "Confirm Sprint Commitment",
+  "question": "Sprint [ID]: [N] stories, [X]d estimated ([Y]% of [Z]d capacity). [Risk level]. Commit?",
+  "header": "Sprint",
   "multiSelect": false,
   "options": [
-    {"label": "Yes, commit plan", "description": "Update status.json and create milestone"},
-    {"label": "No, adjust", "description": "Modify story selection"},
-    {"label": "Cancel", "description": "Exit without changes"}
+    {"label": "Commit sprint plan (Recommended)", "description": "[N] stories across [agents] - [epic summary]. Starts today, ends [date]"},
+    {"label": "Adjust story selection", "description": "Currently at [Y]% capacity - add/remove stories before committing"},
+    {"label": "View dependency risks first", "description": "[risk_count] risks detected: [top risk summary]"},
+    {"label": "Cancel", "description": "Exit without changes to status.json"}
   ]
 }]</parameter>
 </invoke>