moflo 4.6.2 → 4.6.3

package/src/@claude-flow/cli/dist/src/init/moflo-init.js

@@ -130,85 +130,85 @@ function generateConfig(root, force, answers) {
     const detectedExts = extensions.size > 0
         ? [...extensions].sort()
         : ['.ts', '.tsx', '.js', '.jsx'];
-    const yaml = `# MoFlo — Project Configuration
-# Generated by: moflo init
-# Docs: https://github.com/eric-cielo/moflo
-
-project:
-  name: "${projectName}"
-
-# Guidance/knowledge docs to index for semantic search
-guidance:
-  directories:
-${guidanceDirs.map(d => `    - ${d}`).join('\n')}
-  namespace: guidance
-
-# Source directories for code navigation map
-code_map:
-  directories:
-${srcDirs.map(d => `    - ${d}`).join('\n')}
-  extensions: [${detectedExts.map(e => `"${e}"`).join(', ')}]
-  exclude: [node_modules, dist, .next, coverage, build, __pycache__, target, .git]
-  namespace: code-map
-
-# Workflow gates (enforced via Claude Code hooks)
-gates:
-  memory_first: ${gatesEnabled}
-  task_create_first: ${gatesEnabled}
-  context_tracking: ${gatesEnabled}
-
-# Auto-index on session start
-auto_index:
-  guidance: ${answers?.guidance ?? true}
-  code_map: ${answers?.codeMap ?? true}
-
-# Memory backend
-memory:
-  backend: sql.js
-  embedding_model: Xenova/all-MiniLM-L6-v2
-  namespace: default
-
-# Hook toggles (all on by default — disable to slim down)
-hooks:
-  pre_edit: true               # Track file edits for learning
-  post_edit: true              # Record edit outcomes, train neural patterns
-  pre_task: true               # Get agent routing before task spawn
-  post_task: true              # Record task results for learning
-  gate: ${gatesEnabled}                   # Workflow gate enforcement (memory-first, task-create-first)
-  route: true                  # Intelligent task routing on each prompt
-  stop_hook: ${answers?.stopHook ?? true}              # Session-end persistence and metric export
-  session_restore: true        # Restore session state on start
-  notification: true           # Hook into Claude Code notifications
-
-# Status line display (shown at bottom of Claude Code)
-# mode: "single-line" (concise, default) or "dashboard" (full multi-line)
-status_line:
-  enabled: true
-  mode: single-line
-  branding: "MoFlo V4"
-  show_git: true
-  show_model: true
-  show_session: true
-
-# Model preferences (haiku, sonnet, opus)
-models:
-  default: opus        # Model for general tasks
-  research: sonnet     # Model for research/exploration agents
-  review: opus         # Model for code review agents
-  test: sonnet         # Model for test-writing agents
-
-# Intelligent model routing (auto-selects haiku/sonnet/opus per task)
-# When enabled, overrides the static model preferences above
-# by analyzing task complexity and routing to the cheapest capable model.
-model_routing:
-  enabled: false                   # Set to true to enable dynamic routing
-  confidence_threshold: 0.85       # Min confidence before escalating to a more capable model
-  cost_optimization: true          # Prefer cheaper models when confidence is high
-  circuit_breaker: true            # Penalize models that fail repeatedly
-  # Per-agent overrides (set to "inherit" to use routing, or a specific model to pin)
-  # agent_overrides:
-  #   security-architect: opus     # Always use opus for security
-  #   researcher: sonnet           # Pin research to sonnet
+    const yaml = `# MoFlo — Project Configuration
+# Generated by: moflo init
+# Docs: https://github.com/eric-cielo/moflo
+
+project:
+  name: "${projectName}"
+
+# Guidance/knowledge docs to index for semantic search
+guidance:
+  directories:
+${guidanceDirs.map(d => `    - ${d}`).join('\n')}
+  namespace: guidance
+
+# Source directories for code navigation map
+code_map:
+  directories:
+${srcDirs.map(d => `    - ${d}`).join('\n')}
+  extensions: [${detectedExts.map(e => `"${e}"`).join(', ')}]
+  exclude: [node_modules, dist, .next, coverage, build, __pycache__, target, .git]
+  namespace: code-map
+
+# Workflow gates (enforced via Claude Code hooks)
+gates:
+  memory_first: ${gatesEnabled}
+  task_create_first: ${gatesEnabled}
+  context_tracking: ${gatesEnabled}
+
+# Auto-index on session start
+auto_index:
+  guidance: ${answers?.guidance ?? true}
+  code_map: ${answers?.codeMap ?? true}
+
+# Memory backend
+memory:
+  backend: sql.js
+  embedding_model: Xenova/all-MiniLM-L6-v2
+  namespace: default
+
+# Hook toggles (all on by default — disable to slim down)
+hooks:
+  pre_edit: true               # Track file edits for learning
+  post_edit: true              # Record edit outcomes, train neural patterns
+  pre_task: true               # Get agent routing before task spawn
+  post_task: true              # Record task results for learning
+  gate: ${gatesEnabled}                   # Workflow gate enforcement (memory-first, task-create-first)
+  route: true                  # Intelligent task routing on each prompt
+  stop_hook: ${answers?.stopHook ?? true}              # Session-end persistence and metric export
+  session_restore: true        # Restore session state on start
+  notification: true           # Hook into Claude Code notifications
+
+# Status line display (shown at bottom of Claude Code)
+# mode: "single-line" (concise, default) or "dashboard" (full multi-line)
+status_line:
+  enabled: true
+  mode: single-line
+  branding: "MoFlo V4"
+  show_git: true
+  show_model: true
+  show_session: true
+
+# Model preferences (haiku, sonnet, opus)
+models:
+  default: opus        # Model for general tasks
+  research: sonnet     # Model for research/exploration agents
+  review: opus         # Model for code review agents
+  test: sonnet         # Model for test-writing agents
+
+# Intelligent model routing (auto-selects haiku/sonnet/opus per task)
+# When enabled, overrides the static model preferences above
+# by analyzing task complexity and routing to the cheapest capable model.
+model_routing:
+  enabled: false                   # Set to true to enable dynamic routing
+  confidence_threshold: 0.85       # Min confidence before escalating to a more capable model
+  cost_optimization: true          # Prefer cheaper models when confidence is high
+  circuit_breaker: true            # Penalize models that fail repeatedly
+  # Per-agent overrides (set to "inherit" to use routing, or a specific model to pin)
+  # agent_overrides:
+  #   security-architect: opus     # Always use opus for security
+  #   researcher: sonnet           # Pin research to sonnet
 `;
     fs.writeFileSync(configPath, yaml, 'utf-8');
     return { name: 'moflo.yaml', status: 'created', detail: `Detected: ${srcDirs.join(', ')} | ${detectedExts.join(', ')}` };
@@ -401,418 +401,418 @@ function generateSkill(root, force) {
     if (!fs.existsSync(skillDir)) {
         fs.mkdirSync(skillDir, { recursive: true });
     }
-    const skillContent = `---
-name: flo
-description: MoFlo ticket workflow - analyze and execute GitHub issues
-arguments: "[options] <issue-number>"
----
-
-# /flo - MoFlo Ticket Workflow
-
-Research, enhance, and execute GitHub issues automatically.
-
-**Arguments:** $ARGUMENTS
-
-## Usage
-
-\`\`\`
-/flo <issue-number>                   # Full workflow with SWARM (default)
-/flo -e <issue-number>                # Enhance only: research and update ticket, then STOP
-/flo --enhance <issue-number>         # Same as -e
-/flo -r <issue-number>                # Research only: analyze issue, output findings
-/flo --research <issue-number>        # Same as -r
-\`\`\`
-
-Also available as \`/fl\` (shorthand alias).
-
-### Execution Mode (how work is done)
-
-\`\`\`
-/flo 123                              # SWARM mode (default) - multi-agent coordination
-/flo -sw 123                          # SWARM mode (explicit)
-/flo --swarm 123                      # Same as -sw
-/flo -hv 123                          # HIVE-MIND mode - consensus-based coordination
-/flo --hive 123                       # Same as -hv
-/flo -n 123                           # NAKED mode - single Claude, no agents
-/flo --naked 123                      # Same as -n
-\`\`\`
-
-### Epic Handling
-
-\`\`\`
-/flo 42                               # If #42 is an epic, processes all stories sequentially
-\`\`\`
-
-**Epic Detection:** Issues with \`epic\` label or containing \`## Stories\` / \`## Tasks\` sections are automatically detected as epics.
-
-**Sequential Processing:** When an epic is selected:
-1. List all child stories/tasks (from checklist or linked issues)
-2. Process each story **one at a time** in order
-3. Each story goes through the full workflow (research -> enhance -> implement -> test -> PR)
-4. After each story's PR is created, move to the next story
-5. Continue until all stories are complete
-
-### Combined Examples
-
-\`\`\`
-/flo 123                              # Swarm + full workflow (default) - includes ALL tests
-/flo 42                               # If #42 is epic, processes stories sequentially
-/flo -e 123                           # Swarm + enhance only (no implementation)
-/flo -hv -e 123                       # Hive-mind + enhance only
-/flo -n -r 123                        # Naked + research only
-/flo --swarm --enhance 123            # Explicit swarm + enhance only
-/flo -n 123                           # Naked + full workflow (still runs all tests)
-\`\`\`
-
-## SWARM IS MANDATORY BY DEFAULT
-
-Even if a task "looks simple", you MUST use swarm coordination unless
-the user explicitly passes -n/--naked. "Simple" is a trap. Tasks have
-hidden complexity. Swarm catches it.
-
-THE ONLY WAY TO SKIP SWARM: User passes -n or --naked explicitly.
-
-## COMPREHENSIVE TESTING REQUIREMENT
-
-ALL tests MUST pass BEFORE PR creation - NO EXCEPTIONS.
-- Unit Tests: MANDATORY for all new/modified code
-- Integration Tests: MANDATORY for API endpoints and services
-- E2E Tests: MANDATORY for user-facing features
-PR CANNOT BE CREATED until all relevant tests pass.
-
-## Workflow Overview
-
-\`\`\`
-Research -> Enhance -> Execute -> Testing -> Simplify -> PR+Done
-
-Research:    Fetch issue, search memory, read guidance, find files
-Enhance:     Update GitHub issue with tech analysis, affected files, impl plan
-Execute:     Assign self, create branch, implement changes
-Testing:     Unit + Integration + E2E tests (ALL MUST PASS - gate)
-Simplify:    Run /simplify on changed code (gate - must run before PR)
-PR+Done:     Create PR, update issue status, store learnings
-\`\`\`
-
-### Workflow Gates
-
-| Gate | Requirement | Blocked Action |
-|------|-------------|----------------|
-| **Testing Gate** | Unit + Integration + E2E must pass | PR creation |
-| **Simplification Gate** | /simplify must run on changed files | PR creation |
-
-### Execution Mode (applies to all phases)
-
-| Mode | Description |
-|------|-------------|
-| **SWARM** (default) | Multi-agent via Task tool: researcher, coder, tester, reviewer |
-| **HIVE-MIND** (-hv) | Consensus-based coordination for architecture decisions |
-| **NAKED** (-n) | Single Claude, no agent spawning. Only when user explicitly requests. |
-
-## Phase 1: Research (-r or default first step)
-
-### 1.1 Fetch Issue Details
-\`\`\`bash
-gh issue view <issue-number> --json number,title,body,labels,state,assignees,comments,milestone
-\`\`\`
-
-### 1.2 Check Enhancement Status
-Look for \`## Implementation Plan\` marker in issue body.
-- **If present**: Issue already enhanced, skip to execute or confirm
-- **If absent**: Proceed with research and enhancement
-
-### 1.3 Search Memory FIRST
-ALWAYS search memory BEFORE reading guidance or docs files.
-Memory has file paths, context, and patterns - often all you need.
-Only read guidance files if memory search returns zero relevant results.
-
-\`\`\`bash
-npx flo memory search --query "<issue title keywords>" --namespace patterns
-npx flo memory search --query "<domain keywords>" --namespace guidance
-\`\`\`
-
-Or via MCP: \`mcp__claude-flow__memory_search\`
-
-### 1.4 Read Guidance Docs (ONLY if memory insufficient)
-**Only if memory search returned < 3 relevant results**, read guidance files:
-- Bug -> testing patterns, error handling
-- Feature -> domain model, architecture
-- UI -> frontend patterns, components
-
-### 1.5 Research Codebase
-Use Task tool with Explore agent to find:
-- Affected files and their current state
-- Related code and dependencies
-- Existing patterns to follow
-- Test coverage gaps
-
-## Phase 2: Enhance (-e includes research + enhancement)
-
-### 2.1 Build Enhancement
-Compile research into structured enhancement:
-
-**Technical Analysis** - Root cause (bugs) or approach (features), impact, risk factors
-
-**Affected Files** - Files to modify (with line numbers), new files, deletions
-
-**Implementation Plan** - Numbered steps with clear actions, dependencies, decision points
-
-**Test Plan** - Unit tests to add/update, integration tests needed, manual testing checklist
-
-**Estimates** - Complexity (Low/Medium/High), scope (# files, # new tests)
-
-### 2.2 Update GitHub Issue
-\`\`\`bash
-gh issue edit <issue-number> --body "<original body + Technical Analysis + Affected Files + Implementation Plan + Test Plan + Estimates>"
-\`\`\`
-
-### 2.3 Add Enhancement Comment
-\`\`\`bash
-gh issue comment <issue-number> --body "Issue enhanced with implementation plan. Ready for execution."
-\`\`\`
-
-## Phase 3: Execute (default, runs automatically after enhance)
-
-### 3.1 Assign Issue and Update Status
-\`\`\`bash
-gh issue edit <issue-number> --add-assignee @me
-gh issue edit <issue-number> --add-label "in-progress"
-\`\`\`
-
-### 3.2 Create Branch
-\`\`\`bash
-git checkout main && git pull origin main
-git checkout -b <type>/<issue-number>-<short-desc>
-\`\`\`
-Types: \`feature/\`, \`fix/\`, \`refactor/\`, \`docs/\`
-
-### 3.3 Implement
-Follow the implementation plan from the enhanced issue. No prompts - execute all steps.
-
-## Phase 4: Testing (MANDATORY GATE)
-
-This is NOT optional. ALL applicable test types must pass for the change type.
-WORKFLOW STOPS HERE until tests pass. No shortcuts. No exceptions.
-
-### 4.1 Write and Run Tests
-Write unit, integration, and E2E tests as appropriate for the change type.
-Use the project's existing test runner and patterns.
-
-### 4.2 Test Auto-Fix Loop
-If any tests fail, enter the auto-fix loop (max 3 retries OR 10 minutes):
-1. Run all tests
-2. If ALL pass -> proceed to simplification
-3. If ANY fail: analyze failure, fix test or implementation code, retry
-4. If retries exhausted -> STOP and report to user
-
-## Phase 4.5: Code Simplification (MANDATORY)
-
-The built-in /simplify command reviews ALL changed code for:
-- Reuse opportunities and code quality
-- Efficiency improvements
-- Consistency with existing codebase patterns
-- Preserves ALL functionality - no behavior changes
-
-If /simplify makes changes -> re-run tests to confirm nothing broke.
-If re-tests fail -> revert changes, proceed with original code.
-
-## Phase 5: Commit and PR (only after tests pass)
-
-### 5.1 Commit
-\`\`\`bash
-git add <specific files>
-git commit -m "type(scope): description
-
-Closes #<issue-number>
-
-Co-Authored-By: Claude <noreply@anthropic.com>"
-\`\`\`
-
-### 5.2 Create PR
-\`\`\`bash
-git push -u origin <branch-name>
-gh pr create --title "type(scope): description" --body "## Summary
-<brief description>
-
-## Changes
-<bullet list>
-
-## Testing
-- [x] Unit tests pass
-- [x] Integration tests pass
-- [x] E2E tests pass
-- [ ] Manual testing done
-
-Closes #<issue-number>"
-\`\`\`
-
-### 5.3 Update Issue Status
-\`\`\`bash
-gh issue edit <issue-number> --remove-label "in-progress" --add-label "ready-for-review"
-gh issue comment <issue-number> --body "PR created: <pr-url>"
-\`\`\`
-
-## Epic Handling
-
-### Detecting Epics
-
-An issue is an **epic** if:
-1. It has the \`epic\` label, OR
-2. Its body contains \`## Stories\` or \`## Tasks\` sections, OR
-3. It has linked child issues (via \`- [ ] #123\` checklist format)
-
-### Epic Processing Flow
-
-1. DETECT EPIC - Check labels, parse body for ## Stories / ## Tasks, extract issue references
-2. LIST ALL STORIES - Extract from checklist, order top-to-bottom as listed
-3. SEQUENTIAL PROCESSING - For each story: run full /flo workflow, wait for PR, update checklist
-4. COMPLETION - All stories have PRs, epic marked as ready-for-review
-
-ONE STORY AT A TIME - NO PARALLEL STORY EXECUTION.
-Each story must complete (PR created) before starting next.
-
-### Epic Detection Code
-
-\`\`\`javascript
-function isEpic(issue) {
-  if (issue.labels?.some(l => l.name === 'epic')) return true;
-  if (issue.body?.includes('## Stories') || issue.body?.includes('## Tasks')) return true;
-  const linkedIssuePattern = /- \\[[ x]\\] #\\d+/;
-  if (linkedIssuePattern.test(issue.body)) return true;
-  return false;
-}
-
-function extractStories(epicBody) {
-  const stories = [];
-  const pattern = /- \\[[ ]\\] #(\\d+)/g;
-  let match;
-  while ((match = pattern.exec(epicBody)) !== null) {
-    stories.push(parseInt(match[1]));
-  }
-  return stories;
-}
-\`\`\`
-
-## Parse Arguments
-
-\`\`\`javascript
-const args = "$ARGUMENTS".trim().split(/\\s+/);
-let workflowMode = "full";    // full, enhance, research
-let execMode = "swarm";       // swarm (default), hive, naked
-let issueNumber = null;
-
-for (let i = 0; i < args.length; i++) {
-  const arg = args[i];
-
-  // Workflow mode (what to do)
-  if (arg === "-e" || arg === "--enhance") {
-    workflowMode = "enhance";
-  } else if (arg === "-r" || arg === "--research") {
-    workflowMode = "research";
-  }
-
-  // Execution mode (how to do it)
-  else if (arg === "-sw" || arg === "--swarm") {
-    execMode = "swarm";
-  } else if (arg === "-hv" || arg === "--hive") {
-    execMode = "hive";
-  } else if (arg === "-n" || arg === "--naked") {
-    execMode = "naked";
-  }
-
-  // Issue number
-  else if (/^\\d+$/.test(arg)) {
-    issueNumber = arg;
-  }
-}
-
-if (!issueNumber) {
-  throw new Error("Issue number required. Usage: /flo <issue-number>");
-}
-
-// Log execution mode to prevent silent skipping
-console.log("Execution mode: " + execMode.toUpperCase());
-if (execMode === "swarm") {
-  console.log("SWARM MODE: Will spawn agents via Task tool. Do NOT skip this.");
-}
-console.log("TESTING: Unit + Integration + E2E tests REQUIRED before PR.");
-console.log("SIMPLIFY: /simplify command runs on changed code before PR.");
-\`\`\`
-
-## Execution Flow
-
-### Workflow Modes (what to do)
-
-| Mode | Command | Steps | Stops After |
-|------|---------|-------|-------------|
-| **Full** (default) | \`/flo 123\` | Research -> Enhance -> Implement -> Test -> Simplify -> PR | PR created |
-| **Epic** | \`/flo 42\` (epic) | For each story: Full workflow sequentially | All story PRs created |
-| **Enhance** | \`/flo -e 123\` | Research -> Enhance | Issue updated |
-| **Research** | \`/flo -r 123\` | Research | Findings output |
-
-### Execution Modes (how to do it)
-
-| Mode | Flag | Description | When to Use |
-|------|------|-------------|-------------|
-| **Swarm** (DEFAULT) | \`-sw\`, \`--swarm\` | Multi-agent via Task tool | Always, unless explicitly overridden |
-| **Hive-Mind** | \`-hv\`, \`--hive\` | Consensus-based coordination | Architecture decisions, tradeoffs |
-| **Naked** | \`-n\`, \`--naked\` | Single Claude, no agents | User explicitly wants simple mode |
-
-## Execution Mode Details
-
-### SWARM Mode (Default) - ALWAYS USE UNLESS TOLD OTHERWISE
-
-You MUST use the Task tool to spawn agents. No exceptions.
-
-**Swarm spawns these agents via Task tool:**
-- \`researcher\` - Analyzes issue, searches memory, finds patterns
-- \`coder\` - Implements changes following plan
-- \`tester\` - Writes and runs tests
-- \`/simplify\` - Built-in command that reviews changed code before PR
-- \`reviewer\` - Reviews code before PR
-
-**Swarm execution pattern:**
-\`\`\`javascript
-// 1. Create task list FIRST
-TaskCreate({ subject: "Research issue #123", ... })
-TaskCreate({ subject: "Implement changes", ... })
-TaskCreate({ subject: "Test implementation", ... })
-TaskCreate({ subject: "Run /simplify on changed files", ... })
-TaskCreate({ subject: "Review and PR", ... })
-
-// 2. Init swarm
-Bash("npx flo swarm init --topology hierarchical --max-agents 8 --strategy specialized")
-
-// 3. Spawn agents with Task tool (run_in_background: true)
-Task({ prompt: "...", subagent_type: "researcher", run_in_background: true })
-Task({ prompt: "...", subagent_type: "coder", run_in_background: true })
-
-// 4. Wait for results, synthesize, continue
-\`\`\`
-
-### HIVE-MIND Mode (-hv, --hive)
-
-Use for consensus-based decisions:
-- Architecture choices
-- Approach tradeoffs
-- Design decisions with multiple valid options
-
-### NAKED Mode (-n, --naked)
-
-**Only when user explicitly requests.** Single Claude execution without agents.
-- Still uses Task tool for tracking
-- Still creates tasks for visibility
-- Just doesn't spawn multiple agents
-
----
-
-**Full mode executes without prompts.** It will:
-1. Research the issue and codebase
-2. Enhance the GitHub issue with implementation plan
-3. Assign issue to self, add "in-progress" label
-4. Create branch, implement, test
-5. Run /simplify on changed code, re-test if changes made
-6. Commit, create PR, update issue status
-7. Store learnings
+    const skillContent = `---
+name: flo
+description: MoFlo ticket workflow - analyze and execute GitHub issues
+arguments: "[options] <issue-number>"
+---
+
+# /flo - MoFlo Ticket Workflow
+
+Research, enhance, and execute GitHub issues automatically.
+
+**Arguments:** $ARGUMENTS
+
+## Usage
+
+\`\`\`
+/flo <issue-number>                   # Full workflow with SWARM (default)
+/flo -e <issue-number>                # Enhance only: research and update ticket, then STOP
+/flo --enhance <issue-number>         # Same as -e
+/flo -r <issue-number>                # Research only: analyze issue, output findings
+/flo --research <issue-number>        # Same as -r
+\`\`\`
+
+Also available as \`/fl\` (shorthand alias).
+
+### Execution Mode (how work is done)
+
+\`\`\`
+/flo 123                              # SWARM mode (default) - multi-agent coordination
+/flo -sw 123                          # SWARM mode (explicit)
+/flo --swarm 123                      # Same as -sw
+/flo -hv 123                          # HIVE-MIND mode - consensus-based coordination
+/flo --hive 123                       # Same as -hv
+/flo -n 123                           # NAKED mode - single Claude, no agents
+/flo --naked 123                      # Same as -n
+\`\`\`
+
+### Epic Handling
+
+\`\`\`
+/flo 42                               # If #42 is an epic, processes all stories sequentially
+\`\`\`
+
+**Epic Detection:** Issues with \`epic\` label or containing \`## Stories\` / \`## Tasks\` sections are automatically detected as epics.
+
+**Sequential Processing:** When an epic is selected:
+1. List all child stories/tasks (from checklist or linked issues)
+2. Process each story **one at a time** in order
+3. Each story goes through the full workflow (research -> enhance -> implement -> test -> PR)
+4. After each story's PR is created, move to the next story
+5. Continue until all stories are complete
+
+### Combined Examples
+
+\`\`\`
+/flo 123                              # Swarm + full workflow (default) - includes ALL tests
+/flo 42                               # If #42 is epic, processes stories sequentially
+/flo -e 123                           # Swarm + enhance only (no implementation)
+/flo -hv -e 123                       # Hive-mind + enhance only
+/flo -n -r 123                        # Naked + research only
+/flo --swarm --enhance 123            # Explicit swarm + enhance only
+/flo -n 123                           # Naked + full workflow (still runs all tests)
+\`\`\`
+
+## SWARM IS MANDATORY BY DEFAULT
+
+Even if a task "looks simple", you MUST use swarm coordination unless
+the user explicitly passes -n/--naked. "Simple" is a trap. Tasks have
+hidden complexity. Swarm catches it.
+
+THE ONLY WAY TO SKIP SWARM: User passes -n or --naked explicitly.
+
+## COMPREHENSIVE TESTING REQUIREMENT
+
+ALL tests MUST pass BEFORE PR creation - NO EXCEPTIONS.
+- Unit Tests: MANDATORY for all new/modified code
+- Integration Tests: MANDATORY for API endpoints and services
+- E2E Tests: MANDATORY for user-facing features
+PR CANNOT BE CREATED until all relevant tests pass.
+
+## Workflow Overview
+
+\`\`\`
+Research -> Enhance -> Execute -> Testing -> Simplify -> PR+Done
+
+Research:    Fetch issue, search memory, read guidance, find files
+Enhance:     Update GitHub issue with tech analysis, affected files, impl plan
+Execute:     Assign self, create branch, implement changes
+Testing:     Unit + Integration + E2E tests (ALL MUST PASS - gate)
+Simplify:    Run /simplify on changed code (gate - must run before PR)
+PR+Done:     Create PR, update issue status, store learnings
+\`\`\`
+
+### Workflow Gates
+
+| Gate | Requirement | Blocked Action |
+|------|-------------|----------------|
+| **Testing Gate** | Unit + Integration + E2E must pass | PR creation |
+| **Simplification Gate** | /simplify must run on changed files | PR creation |
+
+### Execution Mode (applies to all phases)
+
+| Mode | Description |
+|------|-------------|
+| **SWARM** (default) | Multi-agent via Task tool: researcher, coder, tester, reviewer |
+| **HIVE-MIND** (-hv) | Consensus-based coordination for architecture decisions |
+| **NAKED** (-n) | Single Claude, no agent spawning. Only when user explicitly requests. |
+
+## Phase 1: Research (-r or default first step)
+
+### 1.1 Fetch Issue Details
+\`\`\`bash
+gh issue view <issue-number> --json number,title,body,labels,state,assignees,comments,milestone
+\`\`\`
+
+### 1.2 Check Enhancement Status
+Look for \`## Implementation Plan\` marker in issue body.
+- **If present**: Issue already enhanced, skip to execute or confirm
+- **If absent**: Proceed with research and enhancement
+
+### 1.3 Search Memory FIRST
+ALWAYS search memory BEFORE reading guidance or docs files.
+Memory has file paths, context, and patterns - often all you need.
+Only read guidance files if memory search returns zero relevant results.
+
+\`\`\`bash
+npx flo memory search --query "<issue title keywords>" --namespace patterns
+npx flo memory search --query "<domain keywords>" --namespace guidance
+\`\`\`
+
+Or via MCP: \`mcp__claude-flow__memory_search\`
+
+### 1.4 Read Guidance Docs (ONLY if memory insufficient)
+**Only if memory search returned < 3 relevant results**, read guidance files:
+- Bug -> testing patterns, error handling
+- Feature -> domain model, architecture
+- UI -> frontend patterns, components
+
+### 1.5 Research Codebase
+Use Task tool with Explore agent to find:
+- Affected files and their current state
+- Related code and dependencies
+- Existing patterns to follow
+- Test coverage gaps
+
+## Phase 2: Enhance (-e includes research + enhancement)
+
+### 2.1 Build Enhancement
+Compile research into structured enhancement:
+
+**Technical Analysis** - Root cause (bugs) or approach (features), impact, risk factors
+
+**Affected Files** - Files to modify (with line numbers), new files, deletions
+
+**Implementation Plan** - Numbered steps with clear actions, dependencies, decision points
+
+**Test Plan** - Unit tests to add/update, integration tests needed, manual testing checklist
+
+**Estimates** - Complexity (Low/Medium/High), scope (# files, # new tests)
+
+### 2.2 Update GitHub Issue
+\`\`\`bash
+gh issue edit <issue-number> --body "<original body + Technical Analysis + Affected Files + Implementation Plan + Test Plan + Estimates>"
+\`\`\`
+
+### 2.3 Add Enhancement Comment
+\`\`\`bash
+gh issue comment <issue-number> --body "Issue enhanced with implementation plan. Ready for execution."
+\`\`\`
+
+## Phase 3: Execute (default, runs automatically after enhance)
+
+### 3.1 Assign Issue and Update Status
+\`\`\`bash
+gh issue edit <issue-number> --add-assignee @me
+gh issue edit <issue-number> --add-label "in-progress"
+\`\`\`
+
+### 3.2 Create Branch
+\`\`\`bash
+git checkout main && git pull origin main
+git checkout -b <type>/<issue-number>-<short-desc>
+\`\`\`
+Types: \`feature/\`, \`fix/\`, \`refactor/\`, \`docs/\`
+
+### 3.3 Implement
+Follow the implementation plan from the enhanced issue. No prompts - execute all steps.
+
+## Phase 4: Testing (MANDATORY GATE)
+
+This is NOT optional. ALL applicable test types must pass for the change type.
+WORKFLOW STOPS HERE until tests pass. No shortcuts. No exceptions.
+
+### 4.1 Write and Run Tests
+Write unit, integration, and E2E tests as appropriate for the change type.
+Use the project's existing test runner and patterns.
+
+### 4.2 Test Auto-Fix Loop
+If any tests fail, enter the auto-fix loop (max 3 retries OR 10 minutes):
+1. Run all tests
+2. If ALL pass -> proceed to simplification
+3. If ANY fail: analyze failure, fix test or implementation code, retry
+4. If retries exhausted -> STOP and report to user
+
+## Phase 4.5: Code Simplification (MANDATORY)
+
+The built-in /simplify command reviews ALL changed code for:
+- Reuse opportunities and code quality
+- Efficiency improvements
+- Consistency with existing codebase patterns
+- Preserves ALL functionality - no behavior changes
+
+If /simplify makes changes -> re-run tests to confirm nothing broke.
+If re-tests fail -> revert changes, proceed with original code.
+
+## Phase 5: Commit and PR (only after tests pass)
+
+### 5.1 Commit
+\`\`\`bash
+git add <specific files>
+git commit -m "type(scope): description
+
+Closes #<issue-number>
+
+Co-Authored-By: Claude <noreply@anthropic.com>"
+\`\`\`
+
+### 5.2 Create PR
+\`\`\`bash
+git push -u origin <branch-name>
+gh pr create --title "type(scope): description" --body "## Summary
+<brief description>
+
+## Changes
+<bullet list>
+
+## Testing
+- [x] Unit tests pass
+- [x] Integration tests pass
+- [x] E2E tests pass
+- [ ] Manual testing done
+
+Closes #<issue-number>"
+\`\`\`
+
+### 5.3 Update Issue Status
+\`\`\`bash
+gh issue edit <issue-number> --remove-label "in-progress" --add-label "ready-for-review"
+gh issue comment <issue-number> --body "PR created: <pr-url>"
+\`\`\`
+
+## Epic Handling
+
+### Detecting Epics
+
+An issue is an **epic** if:
+1. It has the \`epic\` label, OR
+2. Its body contains \`## Stories\` or \`## Tasks\` sections, OR
+3. It has linked child issues (via \`- [ ] #123\` checklist format)
+
+### Epic Processing Flow
+
+1. DETECT EPIC - Check labels, parse body for ## Stories / ## Tasks, extract issue references
+2. LIST ALL STORIES - Extract from checklist, order top-to-bottom as listed
+3. SEQUENTIAL PROCESSING - For each story: run full /flo workflow, wait for PR, update checklist
+4. COMPLETION - All stories have PRs, epic marked as ready-for-review
+
+ONE STORY AT A TIME - NO PARALLEL STORY EXECUTION.
+Each story must complete (PR created) before starting next.
+
+### Epic Detection Code
+
+\`\`\`javascript
+function isEpic(issue) {
+  if (issue.labels?.some(l => l.name === 'epic')) return true;
+  if (issue.body?.includes('## Stories') || issue.body?.includes('## Tasks')) return true;
+  const linkedIssuePattern = /- \\[[ x]\\] #\\d+/;
+  if (linkedIssuePattern.test(issue.body)) return true;
+  return false;
+}
+
+function extractStories(epicBody) {
+  const stories = [];
+  const pattern = /- \\[[ ]\\] #(\\d+)/g;
+  let match;
+  while ((match = pattern.exec(epicBody)) !== null) {
+    stories.push(parseInt(match[1]));
+  }
+  return stories;
+}
+\`\`\`
+
+## Parse Arguments
+
+\`\`\`javascript
+const args = "$ARGUMENTS".trim().split(/\\s+/);
+let workflowMode = "full";    // full, enhance, research
+let execMode = "swarm";       // swarm (default), hive, naked
+let issueNumber = null;
+
+for (let i = 0; i < args.length; i++) {
+  const arg = args[i];
+
+  // Workflow mode (what to do)
+  if (arg === "-e" || arg === "--enhance") {
+    workflowMode = "enhance";
+  } else if (arg === "-r" || arg === "--research") {
+    workflowMode = "research";
+  }
+
+  // Execution mode (how to do it)
+  else if (arg === "-sw" || arg === "--swarm") {
+    execMode = "swarm";
+  } else if (arg === "-hv" || arg === "--hive") {
+    execMode = "hive";
+  } else if (arg === "-n" || arg === "--naked") {
+    execMode = "naked";
+  }
+
+  // Issue number
+  else if (/^\\d+$/.test(arg)) {
+    issueNumber = arg;
+  }
+}
+
+if (!issueNumber) {
+  throw new Error("Issue number required. Usage: /flo <issue-number>");
+}
+
+// Log execution mode to prevent silent skipping
+console.log("Execution mode: " + execMode.toUpperCase());
+if (execMode === "swarm") {
+  console.log("SWARM MODE: Will spawn agents via Task tool. Do NOT skip this.");
+}
+console.log("TESTING: Unit + Integration + E2E tests REQUIRED before PR.");
+console.log("SIMPLIFY: /simplify command runs on changed code before PR.");
+\`\`\`
+
+## Execution Flow
+
+### Workflow Modes (what to do)
+
+| Mode | Command | Steps | Stops After |
+|------|---------|-------|-------------|
+| **Full** (default) | \`/flo 123\` | Research -> Enhance -> Implement -> Test -> Simplify -> PR | PR created |
+| **Epic** | \`/flo 42\` (epic) | For each story: Full workflow sequentially | All story PRs created |
+| **Enhance** | \`/flo -e 123\` | Research -> Enhance | Issue updated |
+| **Research** | \`/flo -r 123\` | Research | Findings output |
+
+### Execution Modes (how to do it)
+
+| Mode | Flag | Description | When to Use |
+|------|------|-------------|-------------|
+| **Swarm** (DEFAULT) | \`-sw\`, \`--swarm\` | Multi-agent via Task tool | Always, unless explicitly overridden |
+| **Hive-Mind** | \`-hv\`, \`--hive\` | Consensus-based coordination | Architecture decisions, tradeoffs |
+| **Naked** | \`-n\`, \`--naked\` | Single Claude, no agents | User explicitly wants simple mode |
+
+## Execution Mode Details
+
+### SWARM Mode (Default) - ALWAYS USE UNLESS TOLD OTHERWISE
+
+You MUST use the Task tool to spawn agents. No exceptions.
+
+**Swarm spawns these agents via Task tool:**
+- \`researcher\` - Analyzes issue, searches memory, finds patterns
+- \`coder\` - Implements changes following plan
+- \`tester\` - Writes and runs tests
+- \`/simplify\` - Built-in command that reviews changed code before PR
+- \`reviewer\` - Reviews code before PR
+
+**Swarm execution pattern:**
+\`\`\`javascript
+// 1. Create task list FIRST
+TaskCreate({ subject: "Research issue #123", ... })
+TaskCreate({ subject: "Implement changes", ... })
+TaskCreate({ subject: "Test implementation", ... })
+TaskCreate({ subject: "Run /simplify on changed files", ... })
+TaskCreate({ subject: "Review and PR", ... })
+
+// 2. Init swarm
+Bash("npx flo swarm init --topology hierarchical --max-agents 8 --strategy specialized")
+
+// 3. Spawn agents with Task tool (run_in_background: true)
+Task({ prompt: "...", subagent_type: "researcher", run_in_background: true })
+Task({ prompt: "...", subagent_type: "coder", run_in_background: true })
+
+// 4. Wait for results, synthesize, continue
+\`\`\`
+
+### HIVE-MIND Mode (-hv, --hive)
+
+Use for consensus-based decisions:
+- Architecture choices
+- Approach tradeoffs
+- Design decisions with multiple valid options
+
+### NAKED Mode (-n, --naked)
+
+**Only when user explicitly requests.** Single Claude execution without agents.
+- Still uses Task tool for tracking
+- Still creates tasks for visibility
+- Just doesn't spawn multiple agents
+
+---
+
+**Full mode executes without prompts.** It will:
+1. Research the issue and codebase
+2. Enhance the GitHub issue with implementation plan
+3. Assign issue to self, add "in-progress" label
+4. Create branch, implement, test
+5. Run /simplify on changed code, re-test if changes made
+6. Commit, create PR, update issue status
+7. Store learnings
 `;
     fs.writeFileSync(skillFile, skillContent, 'utf-8');
     // Create /fl alias (same content)
@@ -850,64 +850,64 @@ function generateClaudeMd(root, force) {
             }
         }
     }
-    const mofloSection = `
-${MOFLO_MARKER}
-## MoFlo — AI Agent Orchestration
-
-This project uses [MoFlo](https://github.com/eric-cielo/moflo) for AI-assisted development workflows.
-
-### FIRST ACTION ON EVERY PROMPT: Search Memory
-
-Your first tool call for every new user prompt MUST be a memory search. Do this BEFORE Glob, Grep, Read, or any file exploration.
-
-\`\`\`
-mcp__claude-flow__memory_search  — query: "<task description>", namespace: "guidance" or "patterns" or "code-map"
-\`\`\`
-
-For codebase navigation, search the \`code-map\` namespace first. For patterns and domain knowledge, search \`patterns\` and \`guidance\`.
-
-### Workflow Gates (enforced automatically)
-
-These are enforced by hooks — you cannot bypass them:
-- **Memory-first**: Must search memory before Glob/Grep/Read on guidance files
-- **TaskCreate-first**: Must call TaskCreate before spawning Agent tool
-- **Context tracking**: Session tracked as FRESH → MODERATE → DEPLETED → CRITICAL
-
-### /flo Skill — Issue Execution
-
-Use \`/flo <issue-number>\` (or \`/fl\`) to execute GitHub issues through the full workflow:
-Research → Enhance → Implement → Test → Simplify → PR
-
-### MCP Tools Reference
-
-| Tool | Purpose |
-|------|---------|
-| \`mcp__claude-flow__memory_search\` | Semantic search across indexed knowledge |
-| \`mcp__claude-flow__memory_store\` | Store patterns and decisions |
-| \`mcp__claude-flow__hooks_route\` | Route task to optimal agent type |
-| \`mcp__claude-flow__hooks_pre-task\` | Record task start |
-| \`mcp__claude-flow__hooks_post-task\` | Record task completion for learning |
-
-### Agent Icon Mapping
-
-| Icon | Agent Type | Use For |
-|------|------------|---------|
-| 🔍 | Explore | Research, codebase exploration |
-| 📐 | Plan | Architecture, design |
-| ⚙️ | General | General coding tasks |
-| 🧪 | Test | Writing tests |
-| 🔬 | Analyzer | Code review, analysis |
-| 🔧 | Backend | API implementation |
-
-### Non-Trivial Task Workflow
-
-For any task beyond a single-line fix:
-1. Search memory first (mandatory gate)
-2. Create tasks with TaskCreate (mandatory gate)
-3. Spawn agents in waves (Explore first, then Implement + Test)
-4. Update task status as you go
-5. Store learnings after completion
-${MOFLO_MARKER_END}
+    const mofloSection = `
+${MOFLO_MARKER}
+## MoFlo — AI Agent Orchestration
+
+This project uses [MoFlo](https://github.com/eric-cielo/moflo) for AI-assisted development workflows.
+
+### FIRST ACTION ON EVERY PROMPT: Search Memory
+
+Your first tool call for every new user prompt MUST be a memory search. Do this BEFORE Glob, Grep, Read, or any file exploration.
+
+\`\`\`
+mcp__claude-flow__memory_search  — query: "<task description>", namespace: "guidance" or "patterns" or "code-map"
+\`\`\`
+
+For codebase navigation, search the \`code-map\` namespace first. For patterns and domain knowledge, search \`patterns\` and \`guidance\`.
+
+### Workflow Gates (enforced automatically)
+
+These are enforced by hooks — you cannot bypass them:
+- **Memory-first**: Must search memory before Glob/Grep/Read on guidance files
+- **TaskCreate-first**: Must call TaskCreate before spawning Agent tool
+- **Context tracking**: Session tracked as FRESH → MODERATE → DEPLETED → CRITICAL
+
+### /flo Skill — Issue Execution
+
+Use \`/flo <issue-number>\` (or \`/fl\`) to execute GitHub issues through the full workflow:
+Research → Enhance → Implement → Test → Simplify → PR
+
+### MCP Tools Reference
+
+| Tool | Purpose |
+|------|---------|
+| \`mcp__claude-flow__memory_search\` | Semantic search across indexed knowledge |
+| \`mcp__claude-flow__memory_store\` | Store patterns and decisions |
+| \`mcp__claude-flow__hooks_route\` | Route task to optimal agent type |
+| \`mcp__claude-flow__hooks_pre-task\` | Record task start |
+| \`mcp__claude-flow__hooks_post-task\` | Record task completion for learning |
+
+### Agent Icon Mapping
+
+| Icon | Agent Type | Use For |
+|------|------------|---------|
+| 🔍 | Explore | Research, codebase exploration |
+| 📐 | Plan | Architecture, design |
+| ⚙️ | General | General coding tasks |
+| 🧪 | Test | Writing tests |
+| 🔬 | Analyzer | Code review, analysis |
+| 🔧 | Backend | API implementation |
+
+### Non-Trivial Task Workflow
+
+For any task beyond a single-line fix:
+1. Search memory first (mandatory gate)
+2. Create tasks with TaskCreate (mandatory gate)
+3. Spawn agents in waves (Explore first, then Implement + Test)
+4. Update task status as you go
+5. Store learnings after completion
+${MOFLO_MARKER_END}
 `;
     const finalContent = existing.trimEnd() + '\n' + mofloSection;
     fs.writeFileSync(claudeMdPath, finalContent, 'utf-8');