create-byan-agent 1.0.1 → 1.0.2

package/templates/_bmad/bmb/workflows/byan/delete-agent-workflow.md

@@ -0,0 +1,657 @@
+# BYAN Delete Agent Workflow
+
+**Workflow:** Safe Agent Deletion with Backup & Consequences Warning  
+**Duration:** 5 minutes  
+**Methodology:** Mantra #39 (Consequences) + Safety First
+
+---
+
+## OVERVIEW
+
+Delete Agent workflow safely removes agents with:
+- **Mandatory backup** before deletion
+- **Consequences evaluation** (Mantra #39)
+- **Explicit confirmation** required
+- **Dependency checking**
+- **Rollback capability**
+
+**Key Principle:**
+> "Deletion is permanent. Evaluate consequences, backup everything, confirm explicitly."
+
+**Mantras Applied:**
+- Mantra #39: Évaluation des Conséquences (PRIMARY)
+- Mantra IA-1: Trust But Verify
+- RG-DEL-001: Confirmation explicite
+- RG-DEL-002: Backup obligatoire
+
+---
+
+## SAFETY LEVELS
+
+### 🟢 LOW RISK
+- Agent status: draft
+- Never deployed
+- No dependencies
+- No usage history
+
+→ Simple deletion with backup
+
+### 🟡 MEDIUM RISK
+- Agent status: validated
+- Not deployed but tested
+- Some documentation references
+
+→ Requires confirmation + backup
+
+### 🔴 HIGH RISK
+- Agent status: deployed
+- Actively used in workflows
+- Referenced by other agents
+- Has usage history
+
+→ Requires deprecation period, not immediate deletion
+
+### 🔥 CRITICAL RISK
+- Agent status: deployed
+- Core/critical agent
+- Multiple dependencies
+- High usage frequency
+
+→ **CANNOT DELETE** - Deprecation and migration required
+
+---
+
+## WORKFLOW STEPS
+
+### Step 1: Select Agent to Delete
+
+```
+"⚠️  AGENT DELETION - PERMANENT ACTION
+
+Which agent do you want to delete?
+
+Available agents:
+1. {agent_1} - [{status}] - {usage_info}
+2. {agent_2} - [{status}] - {usage_info}
+3. {agent_3} - [{status}] - {usage_info}
+...
+
+Agent number or name (or 'cancel' to exit): ____
+"
+```
+
+**Load agent and analyze:**
+```python
+agent_spec = AgentSpec.load(agent_id_or_name)
+context = ProjectContext.load(agent_spec.context_id)
+
+# Analyze risk
+risk_level = analyze_deletion_risk(agent_spec)
+dependencies = find_dependencies(agent_spec)
+usage_stats = get_usage_stats(agent_spec)
+```
+
+---
+
+### Step 2: Display Agent Info & Risk Assessment
+
+```
+"════════════════════════════════════════════════════
+AGENT DELETION ANALYSIS
+════════════════════════════════════════════════════
+
+AGENT INFORMATION:
+  Name: {agent_name}
+  Role: {role}
+  Status: {status}
+  Created: {created_at}
+  Last Modified: {updated_at}
+  Last Used: {last_used}
+
+USAGE STATISTICS:
+  Total Invocations: {invocation_count}
+  Active Users: {user_count}
+  Referenced In: {reference_count} files
+  Dependencies: {dependency_count}
+
+RISK ASSESSMENT: {risk_level}
+
+{if LOW RISK 🟢:}
+  This agent can be safely deleted.
+  - Status: draft
+  - No active usage
+  - No dependencies detected
+
+{if MEDIUM RISK 🟡:}
+  This agent has some references.
+  - Status: validated
+  - Referenced in {n} places
+  - Deletion will break: {list_affected}
+
+{if HIGH RISK 🔴:}
+  This agent is actively used!
+  - Status: deployed
+  - {n} active users
+  - {n} dependencies
+  - Immediate deletion NOT RECOMMENDED
+
+{if CRITICAL RISK 🔥:}
+  ⛔ DELETION BLOCKED - This is a critical agent!
+  - Core system component
+  - {n} critical dependencies
+  - Cannot delete without migration plan
+
+════════════════════════════════════════════════════
+"
+```
+
+---
+
+### Step 3: Consequences Evaluation (Mantra #39)
+
+```
+"CONSEQUENCES OF DELETING '{agent_name}':
+
+1. 📊 SCOPE:
+   - Agent will be completely removed
+   - All configurations lost
+   - History lost (unless backed up)
+   Impact: {High/Medium/Low}
+
+2. 💾 DATA:
+   - AgentSpec record deleted
+   - AgentFile(s) deleted
+   - Database entries removed
+   Impact: {High/Medium/Low}
+
+3. 🔧 CODE:
+   - References to this agent will break
+   - Files affected: {list}
+   Impact: {High/Medium/Low}
+
+4. 👥 TEAM:
+   - Users actively using this agent: {count}
+   - Team members to notify: {list}
+   Impact: {High/Medium/Low}
+
+5. 🎯 CLIENTS:
+   - User workflows affected: {workflows}
+   - Breaking change: {yes/no}
+   Impact: {High/Medium/Low}
+
+6. ⚖️ LEGAL:
+   - Compliance impact: {none/review required}
+   Impact: {High/Medium/Low}
+
+7. 🚀 OPERATIONS:
+   - Active processes using agent: {count}
+   - Immediate impact: {yes/no}
+   Impact: {High/Medium/Low}
+
+8. 🔗 DEPENDENCIES:
+   - Agents depending on this one: {list}
+   - External systems affected: {list}
+   Impact: {High/Medium/Low}
+
+9. ⏱️ TIME:
+   - Deletion time: Immediate
+   - Recovery time if mistake: {estimate}
+   - Team notification time: {estimate}
+   Impact: {High/Medium/Low}
+
+10. 🔄 ALTERNATIVES:
+    - Deprecate instead of delete: {recommended/not needed}
+    - Archive and keep backup: {always}
+    - Migrate to replacement agent: {needed/not needed}
+
+OVERALL RISK: {risk_level}
+
+════════════════════════════════════════════════════
+"
+```
+
+---
+
+### Step 4: Dependencies Check
+
+```
+"DEPENDENCY ANALYSIS:
+
+{if no dependencies:}
+  ✅ No dependencies found. Safe to delete.
+
+{if dependencies exist:}
+  ⚠️  Dependencies detected:
+
+  FILES REFERENCING THIS AGENT:
+  1. {file_1}:{line} - {context}
+  2. {file_2}:{line} - {context}
+  ...
+
+  AGENTS DEPENDING ON THIS ONE:
+  - {dependent_agent_1}: {reason}
+  - {dependent_agent_2}: {reason}
+
+  WORKFLOWS USING THIS AGENT:
+  - {workflow_1}
+  - {workflow_2}
+
+  BREAKING IMPACT:
+  - {impact_summary}
+
+  RECOMMENDATION:
+  {if HIGH/CRITICAL:}
+    🛑 DO NOT DELETE immediately.
+    1. Create replacement agent first
+    2. Migrate all dependencies
+    3. Deprecate this agent (30 day notice)
+    4. Delete after migration complete
+
+  {if MEDIUM:}
+    ⚠️  Fix dependencies before deleting:
+    1. Update all referencing files
+    2. Notify users
+    3. Then proceed with deletion
+
+  {if LOW:}
+    ✅ Safe to delete with backup
+"
+```
+
+---
+
+### Step 5: Backup Creation (MANDATORY - RG-DEL-002)
+
+```
+"Creating backup before deletion...
+
+BACKUP STRATEGY:
+1. Export AgentSpec to YAML
+2. Save agent files to archive
+3. Capture full history/metadata
+4. Store in safe location
+
+[▓▓▓▓▓▓▓▓▓▓] 100% - Backup complete
+
+BACKUP CREATED:
+  Location: {backup_path}
+  Format: .tar.gz (compressed)
+  Size: {size}
+  Contains:
+    - agent-spec.yaml
+    - agent-files/
+    - metadata.json
+    - usage-history.log
+
+Backup validated: ✅
+
+To restore this agent later:
+  byan restore {backup_path}
+
+════════════════════════════════════════════════════
+"
+```
+
+**Backup structure:**
+```
+backup-{agent_name}-{timestamp}.tar.gz
+├── agent-spec.yaml          # Full AgentSpec export
+├── agent-files/
+│   ├── copilot.md          # Generated files
+│   ├── vscode.md
+│   └── claude.md
+├── metadata.json           # Timestamps, user, reason
+├── usage-history.log       # Usage statistics
+└── dependencies.json       # Captured dependencies
+```
+
+---
+
+### Step 6: Final Confirmation (RG-DEL-001)
+
+```
+"════════════════════════════════════════════════════
+⚠️  FINAL CONFIRMATION REQUIRED
+════════════════════════════════════════════════════
+
+You are about to PERMANENTLY DELETE:
+  Agent: {agent_name}
+  Role: {role}
+  Status: {status}
+
+This action:
+  ❌ CANNOT be undone (except from backup)
+  ❌ Will break {n} references
+  ❌ Will affect {n} users
+  ✅ Backup created at: {backup_path}
+
+RISK LEVEL: {risk_level}
+
+To confirm deletion, type the agent name exactly: '{agent_name}'
+
+Confirmation: ____
+
+(or type 'cancel' to abort)
+════════════════════════════════════════════════════
+"
+```
+
+**Validation:**
+```python
+def confirm_deletion(user_input, agent_name):
+    if user_input.lower() == 'cancel':
+        return "CANCELLED"
+    
+    if user_input == agent_name:  # Exact match required
+        return "CONFIRMED"
+    
+    return "INVALID"  # Must retype
+```
+
+**If confirmation fails:**
+```
+"❌ Confirmation failed. Agent name must match exactly.
+
+Expected: '{agent_name}'
+You typed: '{user_input}'
+
+Try again or type 'cancel': ____
+"
+```
+
+---
+
+### Step 7: Execute Deletion
+
+**If confirmed:**
+```python
+def delete_agent(agent_spec, backup):
+    try:
+        # 1. Delete agent files
+        agent_files = AgentFile.find_by_agent(agent_spec.id)
+        for file in agent_files:
+            file.delete()
+            log_deletion(file, backup)
+        
+        # 2. Delete AgentSpec
+        agent_spec.delete()
+        log_deletion(agent_spec, backup)
+        
+        # 3. Update project context (remove agent reference)
+        context = ProjectContext.load(agent_spec.context_id)
+        context.remove_agent_reference(agent_spec.id)
+        context.save()
+        
+        # 4. Create deletion record (audit trail)
+        DeletionRecord.create(
+            agent_name=agent_spec.agent_name,
+            deleted_by=current_user,
+            deleted_at=now(),
+            backup_location=backup.path,
+            reason=deletion_reason,
+            consequences=consequences_evaluated
+        )
+        
+        return DeleteSuccess(backup=backup)
+    
+    except Exception as e:
+        # Rollback if anything fails
+        rollback_deletion(backup)
+        return DeleteFailed(error=e)
+```
+
+**Progress:**
+```
+"Deleting agent...
+
+[▓▓▓░░░░░░░] 30% - Removing agent files...
+[▓▓▓▓▓▓░░░░] 60% - Deleting AgentSpec...
+[▓▓▓▓▓▓▓▓░░] 80% - Updating project context...
+[▓▓▓▓▓▓▓▓▓▓] 100% - Creating audit record...
+
+✅ Deletion complete.
+"
+```
+
+---
+
+### Step 8: Post-Deletion Summary
+
+```
+"════════════════════════════════════════════════════
+DELETION COMPLETE
+════════════════════════════════════════════════════
+
+Agent '{agent_name}' has been permanently deleted.
+
+WHAT WAS DELETED:
+  ❌ AgentSpec record
+  ❌ {n} agent file(s)
+  ❌ Agent configurations
+  ❌ Menu references
+
+WHAT WAS PRESERVED:
+  ✅ Full backup: {backup_path}
+  ✅ Deletion audit record
+  ✅ Usage history (in backup)
+
+NEXT STEPS:
+
+{if dependencies existed:}
+  ⚠️  ACTION REQUIRED:
+  1. Fix broken references:
+     {list_files_to_fix}
+  2. Notify affected users:
+     {list_users}
+  3. Update documentation
+
+{if no dependencies:}
+  ✅ No further action needed.
+
+RESTORATION:
+  To restore this agent from backup:
+    byan restore {backup_path}
+
+  Backup will be kept for {retention_period} days.
+
+AUDIT TRAIL:
+  Deletion recorded in: {audit_log_path}
+  Deleted by: {user}
+  Timestamp: {timestamp}
+  Reason: {reason}
+
+════════════════════════════════════════════════════
+"
+```
+
+---
+
+## ALTERNATIVE: DEPRECATION (Recommended for Deployed Agents)
+
+**If agent is HIGH or CRITICAL risk:**
+
+```
+"🛑 DELETION NOT RECOMMENDED
+
+This agent is actively used. Instead of deleting immediately,
+I recommend DEPRECATION:
+
+DEPRECATION WORKFLOW:
+1. Mark agent as 'deprecated' (status change)
+2. Add deprecation notice (30-60 day warning)
+3. Notify all users
+4. Provide migration path to replacement
+5. Monitor usage decline
+6. Delete after usage reaches zero
+
+Start deprecation workflow? (yes/cancel): ____
+"
+```
+
+**Deprecation process:**
+```python
+def deprecate_agent(agent_spec, replacement=None):
+    # 1. Change status
+    agent_spec.status = "deprecated"
+    agent_spec.deprecation_date = now()
+    agent_spec.deprecation_deadline = now() + timedelta(days=30)
+    
+    if replacement:
+        agent_spec.replacement_agent = replacement.id
+    
+    # 2. Add deprecation warning to agent
+    agent_spec.add_deprecation_notice(
+        message=f"This agent is deprecated and will be removed on {deadline}",
+        replacement=replacement.agent_name if replacement else None,
+        migration_guide=generate_migration_guide(agent_spec, replacement)
+    )
+    
+    # 3. Notify users
+    notify_users(
+        agent=agent_spec,
+        message="Agent deprecated - action required",
+        deadline=agent_spec.deprecation_deadline
+    )
+    
+    agent_spec.save()
+```
+
+---
+
+## RESTORE FROM BACKUP
+
+**If user made a mistake:**
+
+```
+"Restore agent from backup:
+
+  byan restore {backup_path}
+
+This will:
+  ✅ Recreate AgentSpec
+  ✅ Restore all agent files
+  ✅ Restore configurations
+  ✅ Preserve original timestamps
+
+Note: Any changes made after backup will be lost.
+"
+```
+
+```python
+def restore_agent(backup_path):
+    backup = Backup.load(backup_path)
+    
+    # 1. Validate backup integrity
+    if not backup.is_valid():
+        raise BackupCorruptedError()
+    
+    # 2. Check if agent name conflicts
+    if AgentSpec.exists(backup.agent_name):
+        raise AgentAlreadyExistsError()
+    
+    # 3. Restore AgentSpec
+    agent_spec = AgentSpec.from_backup(backup)
+    agent_spec.save()
+    
+    # 4. Restore agent files
+    for file_data in backup.files:
+        agent_file = AgentFile.from_backup(file_data)
+        agent_file.save()
+    
+    # 5. Update project context
+    context = ProjectContext.load(agent_spec.context_id)
+    context.add_agent_reference(agent_spec.id)
+    context.save()
+    
+    return RestoreSuccess(agent_spec)
+```
+
+---
+
+## ANTI-PATTERNS
+
+**NEVER do these:**
+
+❌ Delete without backup
+❌ Delete deployed agents immediately
+❌ Delete without checking dependencies
+❌ Delete without team notification
+❌ Delete without explicit confirmation
+❌ Delete without consequence evaluation
+❌ Delete without audit trail
+
+**ALWAYS do these:**
+
+✅ Create backup first (RG-DEL-002)
+✅ Evaluate consequences (Mantra #39)
+✅ Check dependencies thoroughly
+✅ Require explicit confirmation (RG-DEL-001)
+✅ Notify affected users
+✅ Prefer deprecation for deployed agents
+✅ Keep audit trail
+✅ Test restore procedure
+
+---
+
+## CANCELLATION
+
+**At any point, user can cancel:**
+
+```
+"Deletion cancelled.
+
+No changes were made. Agent '{agent_name}' is still active.
+
+{if backup was created:}
+  Note: Backup was created but not used.
+  Location: {backup_path}
+  You can delete this backup manually if not needed.
+
+Return to menu? (yes): ____
+"
+```
+
+---
+
+## SUCCESS CRITERIA
+
+✅ Risk level assessed
+✅ Dependencies checked
+✅ Consequences evaluated (10 dimensions)
+✅ Backup created and validated (RG-DEL-002)
+✅ Explicit confirmation obtained (RG-DEL-001)
+✅ Agent deleted successfully
+✅ Audit trail created
+✅ User notified of next steps
+✅ Restore instructions provided
+
+**For HIGH/CRITICAL risk agents:**
+✅ Deprecation recommended instead of deletion
+✅ Migration path provided
+✅ Users notified of deprecation
+
+---
+
+## COMPLETION
+
+```
+"Delete Agent workflow complete.
+
+RESULT: {SUCCESS / CANCELLED / FAILED}
+
+{if SUCCESS:}
+  Agent '{agent_name}' has been deleted.
+  Backup: {backup_path}
+  Audit: {audit_path}
+
+{if CANCELLED:}
+  No changes made.
+
+{if FAILED:}
+  Error: {error_message}
+  Agent NOT deleted.
+  Rollback successful.
+
+Session summary: {output_folder}/delete-agent-{agent_name}-{timestamp}.md
+"
+```