Merge 4d48b0dbe1 into ed0defbe08

* fix: normalize workflow manifest schema

* fix: escape workflow manifest values safely

---------

Co-authored-by: Brian <bmadcode@gmail.com>

docs/ring-of-fire-sessions.md

@@ -0,0 +1,256 @@
+# BMad Method PR #1: Ring of Fire (ROF) Sessions
+
+**Feature Type**: Core workflow enhancement
+**Status**: Draft for community review
+**Origin**: tellingCube project (masemIT e.U.)
+**Author**: Mario Semper (@sempre)
+**Date**: 2025-11-23
+
+---
+
+## Summary
+
+**Ring of Fire (ROF) Sessions** enable multi-agent collaborative sessions that run in parallel to the user's main workflow, allowing users to delegate complex multi-perspective analysis while continuing other work.
+
+---
+
+## Problem Statement
+
+Current BMad Method requires **sequential agent interaction**. When users need multiple agents to collaborate on a complex topic, they must:
+- Manually orchestrate each agent conversation
+- Stay in the loop for every exchange
+- Wait for sequential responses before proceeding
+- Context-switch constantly between tasks
+
+This creates **bottlenecks** and prevents **parallel work streams**.
+
+---
+
+## Proposed Solution: Ring of Fire Sessions
+
+A new command pattern that enables **scoped multi-agent collaboration sessions** that run while the user continues other work.
+
+### Command Syntax
+
+```bash
+*rof "<topic>" --agents <agent-list> [--report brief|detailed|live]
+```
+
+### Example Usage
+
+```bash
+*rof "API Refactoring Strategy" --agents dev,architect,qa --report brief
+```
+
+**What happens**:
+1. Dev, Architect, and QA agents enter a collaborative session
+2. They analyze the topic together (code review, design discussion, testing concerns)
+3. When agents need tool access (read files, run commands), they request user approval
+4. User continues working on other tasks in parallel
+5. Session ends with consolidated report (brief: just recommendations, detailed: full transcript)
+
+---
+
+## Key Features
+
+### 1. User-Controlled Scope
+- **Small**: 2 agents, 5-minute quick discussion
+- **Large**: 10 agents, 2-hour deep analysis
+- User decides granularity based on complexity
+
+### 2. Approval-Gated Tool Access
+- Agents can **discuss** freely within the session
+- When agents need **tools** (read files, execute commands, make changes), they:
+  - Pause the session
+  - Request user approval
+  - Resume after user decision
+
+**Why**: Maintains user control, prevents runaway agent actions
+
+### 3. Flexible Reporting
+
+| Mode | Description | Use Case |
+|------|-------------|----------|
+| `brief` | Final recommendations only | "Just tell me what to do" |
+| `detailed` | Full transcript + recommendations | "Show me the reasoning" |
+| `live` | Real-time updates as agents discuss | "I want to observe" |
+
+**Default**: `brief` with Q&A available
+
+### 4. Parallel Workflows
+- User works on **Task A** while ROF session tackles **Task B**
+- No context-switching overhead
+- Efficient use of time
+
+---
+
+## Use Cases
+
+### 1. Architecture Reviews
+```bash
+*rof "Evaluate microservices vs monolith for new feature" --agents architect,dev,qa
+```
+**Agents collaborate on**: Design trade-offs, implementation complexity, testing implications
+
+### 2. Code Refactoring
+```bash
+*rof "Refactor authentication module" --agents dev,architect --report detailed
+```
+**Agents collaborate on**: Current code analysis, refactoring approach, migration strategy
+
+### 3. Feature Planning
+```bash
+*rof "Plan user notifications feature" --agents pm,ux,dev --report brief
+```
+**Agents collaborate on**: Requirements, UX flow, technical feasibility, timeline
+
+### 4. Quality Gates
+```bash
+*rof "Investigate test failures in CI/CD" --agents qa,dev --report live
+```
+**Agents collaborate on**: Root cause analysis, fix recommendations, regression prevention
+
+### 5. Documentation Sprints
+```bash
+*rof "Document API endpoints" --agents dev,pm,ux
+```
+**Agents collaborate on**: Technical accuracy, user-friendly examples, completeness
+
+---
+
+## User Experience Flow
+
+```mermaid
+sequenceDiagram
+    User->>River: *rof "Topic" --agents dev,architect
+    River->>Dev: Join ROF session
+    River->>Architect: Join ROF session
+    River->>User: Session started, continue your work
+
+    Dev->>Architect: Discuss approach
+    Architect->>Dev: Suggest alternatives
+
+    Dev->>User: Need to read auth.ts - approve?
+    User->>Dev: Approved
+    Dev->>Architect: After reading file...
+
+    Architect->>Dev: Recommendation
+    Dev->>River: Session complete
+    River->>User: Brief report: [Recommendations]
+```
+
+---
+
+## Implementation Considerations
+
+### Technical Requirements
+- **Session state management**: Track active ROF sessions, participating agents
+- **Agent context sharing**: Agents share knowledge within session scope
+- **User approval workflow**: Clear prompt for tool requests
+- **Report generation**: Brief/detailed/live output formatting
+- **Workflow integration**: Link ROF findings to existing workflow plans/todos
+
+### Open Questions for Community
+
+1. **Integration**: Core BMad feature or plugin/extension?
+2. **Concurrency**: How to handle file conflicts if multiple agents want to edit?
+3. **Cost Model**: Guidance for LLM call budgeting with multiple agents?
+4. **Session Limits**: Recommended max agents/duration?
+5. **Agent Communication**: Free-form discussion or structured turn-taking?
+
+---
+
+## Real-World Validation
+
+**Origin Project**: tellingCube (BI dashboard, masemIT e.U.)
+
+**Validation Scenario**:
+- **Topic**: "Next steps for tellingCube after validation test"
+- **Agents**: River (orchestrator), Mary (analyst), Winston (architect)
+- **Report Mode**: Brief
+- **Outcome**: Successfully analyzed post-validation roadmap with 3 scenarios (GO/CHANGE/NO-GO), delivered consolidated recommendations in 5 minutes
+
+**User Feedback (Mario Semper)**:
+> "This is exactly what I needed - I wanted multiple perspectives without having to orchestrate every conversation. The brief report gave me actionable next steps immediately."
+
+**Documentation**: `docs/_masemIT/readme.md` in tellingCube repository
+
+---
+
+## Proposed Documentation Structure
+
+```
+.bmad-core/
+  features/
+    ring-of-fire.md              # Feature specification
+
+docs/
+  guides/
+    using-rof-sessions.md        # User guide with examples
+
+  architecture/
+    agent-collaboration.md       # Technical design
+    rof-session-management.md    # State handling approach
+```
+
+---
+
+## Benefits
+
+✅ **Unlocks parallel workflows** - User productivity gains
+✅ **Reduces context-switching** - Cognitive load reduction
+✅ **Enables complex analysis** - Multi-perspective insights
+✅ **Maintains user control** - Approval gates for tools
+✅ **Scales flexibly** - From quick checks to deep dives
+
+---
+
+## Comparison to Existing Patterns
+
+| Feature | Standard Agent Use | ROF Session |
+|---------|-------------------|-------------|
+| Agent collaboration | Sequential (one at a time) | Parallel (multiple simultaneously) |
+| User involvement | Required for every exchange | Only for approvals |
+| Parallel work | No (user waits) | Yes (user continues tasks) |
+| Output | Chat transcript | Consolidated report |
+| Use case | Single-perspective tasks | Multi-perspective analysis |
+
+---
+
+## Next Steps
+
+1. **Community feedback** on approach and open questions
+2. **Technical design** refinement (state management, agent communication)
+3. **Prototype implementation** in BMad core or as extension
+4. **Beta testing** with real projects (beyond tellingCube)
+5. **Documentation** completion with examples
+
+---
+
+## Alternatives Considered
+
+### Alt 1: "Breakout Session"
+- **Pros**: Clear meeting metaphor
+- **Cons**: Less evocative, doesn't convey "continuous collaborative space"
+
+### Alt 2: "Agent Huddle"
+- **Pros**: Short, casual
+- **Cons**: Implies quick/informal only
+
+### Alt 3: "Lagerfeuer" (original German name)
+- **Pros**: Warm, campfire metaphor
+- **Cons**: Poor i18n, hard to pronounce/remember for non-German speakers
+
+**Chosen**: **Ring of Fire** - evokes continuous collaboration circle, internationally understood, memorable, shortcut "ROF" works well
+
+---
+
+## References
+
+- **Source Project**: tellingCube (https://github.com/masemIT/telling-cube) [if public]
+- **Documentation**: `docs/_masemIT/readme.md`
+- **Discussion**: [Link to BMad community discussion if applicable]
+
+---
+
+**Contribution ready for review.** Feedback welcome! 🔥

src/modules/bmgd/workflows/4-production/code-review/instructions.md

@@ -330,7 +330,7 @@ Review was saved to story file, but sprint-status.yaml may be out of sync.
       <action>All action items are included in the standalone review report</action>
       <ask if="action items exist">Would you like me to create tracking items for these action items? (backlog/tasks)</ask>
       <action if="user confirms">
-        If {{backlog_file}} does not exist, copy {installed_path}/backlog_template.md to {{backlog_file}} location.
+        If {{backlog_file}} does not exist, copy {installed_path}/backlog-template.md to {{backlog_file}} location.
         Append a row per action item with Date={{date}}, Story="Ad-Hoc Review", Epic="N/A", Type, Severity, Owner (or "TBD"), Status="Open", Notes with file refs and context.
       </action>
     </check>
@@ -342,7 +342,7 @@ Review was saved to story file, but sprint-status.yaml may be out of sync.
         Append under the story's "Tasks / Subtasks" a new subsection titled "Review Follow-ups (AI)", adding each item as an unchecked checkbox in imperative form, prefixed with "[AI-Review]" and severity. Example: "- [ ] [AI-Review][High] Add input validation on server route /api/x (AC #2)".
       </action>
       <action>
-        If {{backlog_file}} does not exist, copy {installed_path}/backlog_template.md to {{backlog_file}} location.
+        If {{backlog_file}} does not exist, copy {installed_path}/backlog-template.md to {{backlog_file}} location.
         Append a row per action item with Date={{date}}, Story={{epic_num}}.{{story_num}}, Epic={{epic_num}}, Type, Severity, Owner (or "TBD"), Status="Open", Notes with short context and file refs.
       </action>
       <action>

src/modules/bmm/agents/dev.agent.yaml

@@ -24,7 +24,7 @@ agent:
 
   critical_actions:
     - "READ the entire story file BEFORE any implementation - tasks/subtasks sequence is your authoritative implementation guide"
-    - "Load project_context.md if available for coding standards only - never let it override story requirements"
+    - "Load project-context.md if available for coding standards only - never let it override story requirements"
     - "Execute tasks/subtasks IN ORDER as written in story file - no skipping, no reordering, no doing what you want"
     - "For each task/subtask: follow red-green-refactor cycle - write failing test first, then implementation"
     - "Mark task/subtask [x] ONLY when both implementation AND tests are complete and passing"

src/modules/bmm/workflows/2-plan-workflows/prd/workflow.md

@@ -1,6 +1,6 @@
 ---
 name: create-prd
-description: Creates a comprehensive PRDs through collaborative step-by-step discovery between two product managers working as peers.
+description: Creates a comprehensive PRD through collaborative step-by-step discovery between two product managers working as peers.
 main_config: '{project-root}/.bmad/bmm/config.yaml'
 web_bundle: true
 ---

src/modules/bmm/workflows/3-solutioning/architecture/steps/step-01-init.md

@@ -94,7 +94,7 @@ Discover and load context documents using smart discovery:
 
 **Project Context Rules (Critical for AI Agents):**
 
-1. Check for project context file: `**/project_context.md`
+1. Check for project context file: `**/project-context.md`
 2. If exists: Load COMPLETE file contents - this contains critical rules for AI agents
 3. Add to frontmatter `hasProjectContext: true` and track file path
 4. Report to user: "Found existing project context with {number_of_rules} agent rules"

src/modules/bmm/workflows/3-solutioning/architecture/steps/step-08-complete.md

@@ -280,7 +280,7 @@ Your architecture will ensure consistent, high-quality implementation across all
 
 **💡 Optional Enhancement: Project Context File**
 
-Would you like to create a `project_context.md` file? This is a concise, optimized guide for AI agents that captures:
+Would you like to create a `project-context.md` file? This is a concise, optimized guide for AI agents that captures:
 
 - Critical language and framework rules they might miss
 - Specific patterns and conventions for your project
@@ -310,7 +310,7 @@ This will help ensure consistent implementation by capturing:
 - Testing and quality standards
 - Anti-patterns to avoid
 
-The workflow will collaborate with you to create an optimized `project_context.md` file that AI agents will read before implementing any code."
+The workflow will collaborate with you to create an optimized `project-context.md` file that AI agents will read before implementing any code."
 
 **Execute the Generate Project Context workflow:**
 

src/modules/bmm/workflows/4-implementation/code-review/instructions.xml

@@ -217,7 +217,7 @@
       **Issues Fixed:** {{fixed_count}}
       **Action Items Created:** {{action_count}}
 
-      {{#if new_status == "done"}}Story is ready for next work!{{else}}Address the action items and continue development.{{/if}}
+      {{#if new_status == "done"}}Code review complete!{{else}}Address the action items and continue development.{{/if}}
     </output>
   </step>
 

src/modules/bmm/workflows/4-implementation/dev-story/checklist.md

@@ -35,7 +35,7 @@ validation-rules:
 - [ ] **Acceptance Criteria Satisfaction:** Implementation satisfies EVERY Acceptance Criterion in the story
 - [ ] **No Ambiguous Implementation:** Clear, unambiguous implementation that meets story requirements
 - [ ] **Edge Cases Handled:** Error conditions and edge cases appropriately addressed
-- [ ] **Dependencies Within Scope:** Only uses dependencies specified in story or project_context.md
+- [ ] **Dependencies Within Scope:** Only uses dependencies specified in story or project-context.md
 
 ## 🧪 Testing & Quality Assurance
 

src/modules/bmm/workflows/generate-project-context/steps/step-01-discover.md

@@ -33,7 +33,7 @@ Discover the project's technology stack, existing patterns, and critical impleme
 
 First, check if project context already exists:
 
-- Look for file at `{output_folder}/project_context.md`
+- Look for file at `{output_folder}/project-context.md`
 - If exists: Read complete file to understand existing rules
 - Present to user: "Found existing project context with {number_of_sections} sections. Would you like to update this or create a new one?"
 
@@ -122,7 +122,7 @@ Based on discovery, create or update the context document:
 
 #### A. Fresh Document Setup (if no existing context)
 
-Copy template from `{installed_path}/project-context-template.md` to `{output_folder}/project_context.md`
+Copy template from `{installed_path}/project-context-template.md` to `{output_folder}/project-context.md`
 Initialize frontmatter with:
 
 ```yaml

src/modules/bmm/workflows/generate-project-context/steps/step-02-generate.md

@@ -288,7 +288,7 @@ After each category, show the generated rules and present choices:
 
 ## APPEND TO PROJECT CONTEXT:
 
-When user selects 'C' for a category, append the content directly to `{output_folder}/project_context.md` using the structure from step 8.
+When user selects 'C' for a category, append the content directly to `{output_folder}/project-context.md` using the structure from step 8.
 
 ## SUCCESS METRICS:
 

src/modules/bmm/workflows/generate-project-context/steps/step-03-complete.md

@@ -134,7 +134,7 @@ Based on user skill level, present the completion:
 **Expert Mode:**
 "Project context complete. Optimized for LLM consumption with {{rule_count}} critical rules across {{section_count}} sections.
 
-File saved to: `{output_folder}/project_context.md`
+File saved to: `{output_folder}/project-context.md`
 
 Ready for AI agent integration."
 
@@ -227,7 +227,7 @@ Present final completion to user:
 "✅ **Project Context Generation Complete!**
 
 Your optimized project context file is ready at:
-`{output_folder}/project_context.md`
+`{output_folder}/project-context.md`
 
 **📊 Context Summary:**
 

src/modules/bmm/workflows/generate-project-context/workflow.md

@@ -1,11 +1,11 @@
 ---
 name: generate-project-context
-description: Creates a concise project_context.md file with critical rules and patterns that AI agents must follow when implementing code. Optimized for LLM context efficiency.
+description: Creates a concise project-context.md file with critical rules and patterns that AI agents must follow when implementing code. Optimized for LLM context efficiency.
 ---
 
 # Generate Project Context Workflow
 
-**Goal:** Create a concise, optimized `project_context.md` file containing critical rules, patterns, and guidelines that AI agents must follow when implementing code. This file focuses on unobvious details that LLMs need to be reminded of.
+**Goal:** Create a concise, optimized `project-context.md` file containing critical rules, patterns, and guidelines that AI agents must follow when implementing code. This file focuses on unobvious details that LLMs need to be reminded of.
 
 **Your Role:** You are a technical facilitator working with a peer to capture the essential implementation rules that will ensure consistent, high-quality code generation across all AI agents working on the project.
 
@@ -37,7 +37,7 @@ Load config from `{project-root}/.bmad/bmm/config.yaml` and resolve:
 
 - `installed_path` = `{project-root}/.bmad/bmm/workflows/generate-project-context`
 - `template_path` = `{installed_path}/project-context-template.md`
-- `output_file` = `{output_folder}/project_context.md`
+- `output_file` = `{output_folder}/project-context.md`
 
 ---
 

tools/cli/installers/lib/core/config-collector.js

@@ -248,14 +248,21 @@ class ConfigCollector {
     const configKeys = Object.keys(moduleConfig).filter((key) => key !== 'prompt');
     const existingKeys = this.existingConfig && this.existingConfig[moduleName] ? Object.keys(this.existingConfig[moduleName]) : [];
 
+    // Find new interactive fields (with prompt)
     const newKeys = configKeys.filter((key) => {
       const item = moduleConfig[key];
       // Check if it's a config item and doesn't exist in existing config
       return item && typeof item === 'object' && item.prompt && !existingKeys.includes(key);
     });
 
-    // If in silent mode and no new keys, use existing config and skip prompts
-    if (silentMode && newKeys.length === 0) {
+    // Find new static fields (without prompt, just result)
+    const newStaticKeys = configKeys.filter((key) => {
+      const item = moduleConfig[key];
+      return item && typeof item === 'object' && !item.prompt && item.result && !existingKeys.includes(key);
+    });
+
+    // If in silent mode and no new keys (neither interactive nor static), use existing config and skip prompts
+    if (silentMode && newKeys.length === 0 && newStaticKeys.length === 0) {
       if (this.existingConfig && this.existingConfig[moduleName]) {
         if (!this.collectedConfig[moduleName]) {
           this.collectedConfig[moduleName] = {};
@@ -294,9 +301,12 @@ class ConfigCollector {
       return false; // No new fields
     }
 
-    // If we have new fields, build questions first
-    if (newKeys.length > 0) {
+    // If we have new fields (interactive or static), process them
+    if (newKeys.length > 0 || newStaticKeys.length > 0) {
       const questions = [];
+      const staticAnswers = {};
+
+      // Build questions for interactive fields
       for (const key of newKeys) {
         const item = moduleConfig[key];
         const question = await this.buildQuestion(moduleName, key, item, moduleConfig);
@@ -305,39 +315,50 @@ class ConfigCollector {
         }
       }
 
+      // Prepare static answers (no prompt, just result)
+      for (const key of newStaticKeys) {
+        staticAnswers[`${moduleName}_${key}`] = undefined;
+      }
+
+      // Collect all answers (static + prompted)
+      let allAnswers = { ...staticAnswers };
+
       if (questions.length > 0) {
         // Only show header if we actually have questions
         CLIUtils.displayModuleConfigHeader(moduleName, moduleConfig.header, moduleConfig.subheader);
         console.log(); // Line break before questions
-        const answers = await inquirer.prompt(questions);
+        const promptedAnswers = await inquirer.prompt(questions);
 
-        // Store answers for cross-referencing
-        Object.assign(this.allAnswers, answers);
-
-        // Process answers and build result values
-        for (const key of Object.keys(answers)) {
-          const originalKey = key.replace(`${moduleName}_`, '');
-          const item = moduleConfig[originalKey];
-          const value = answers[key];
-
-          let result;
-          if (Array.isArray(value)) {
-            result = value;
-          } else if (item.result) {
-            result = this.processResultTemplate(item.result, value);
-          } else {
-            result = value;
-          }
-
-          if (!this.collectedConfig[moduleName]) {
-            this.collectedConfig[moduleName] = {};
-          }
-          this.collectedConfig[moduleName][originalKey] = result;
-        }
-      } else {
-        // New keys exist but no questions generated - show no config message
+        // Merge prompted answers with static answers
+        Object.assign(allAnswers, promptedAnswers);
+      } else if (newStaticKeys.length > 0) {
+        // Only static fields, no questions - show no config message
         CLIUtils.displayModuleNoConfig(moduleName, moduleConfig.header, moduleConfig.subheader);
       }
+
+      // Store all answers for cross-referencing
+      Object.assign(this.allAnswers, allAnswers);
+
+      // Process all answers (both static and prompted)
+      for (const key of Object.keys(allAnswers)) {
+        const originalKey = key.replace(`${moduleName}_`, '');
+        const item = moduleConfig[originalKey];
+        const value = allAnswers[key];
+
+        let result;
+        if (Array.isArray(value)) {
+          result = value;
+        } else if (item.result) {
+          result = this.processResultTemplate(item.result, value);
+        } else {
+          result = value;
+        }
+
+        if (!this.collectedConfig[moduleName]) {
+          this.collectedConfig[moduleName] = {};
+        }
+        this.collectedConfig[moduleName][originalKey] = result;
+      }
     }
 
     // Copy over existing values for fields that weren't prompted
@@ -353,7 +374,7 @@ class ConfigCollector {
       }
     }
 
-    return newKeys.length > 0; // Return true if we prompted for new fields
+    return newKeys.length > 0 || newStaticKeys.length > 0; // Return true if we had any new fields (interactive or static)
   }
 
   /**
@@ -501,30 +522,52 @@ class ConfigCollector {
 
     // Process each config item
     const questions = [];
+    const staticAnswers = {};
     const configKeys = Object.keys(moduleConfig).filter((key) => key !== 'prompt');
 
     for (const key of configKeys) {
       const item = moduleConfig[key];
 
       // Skip if not a config object
-      if (!item || typeof item !== 'object' || !item.prompt) {
+      if (!item || typeof item !== 'object') {
         continue;
       }
 
-      const question = await this.buildQuestion(moduleName, key, item, moduleConfig);
-      if (question) {
-        questions.push(question);
+      // Handle static values (no prompt, just result)
+      if (!item.prompt && item.result) {
+        // Add to static answers with a marker value
+        staticAnswers[`${moduleName}_${key}`] = undefined;
+        continue;
+      }
+
+      // Handle interactive values (with prompt)
+      if (item.prompt) {
+        const question = await this.buildQuestion(moduleName, key, item, moduleConfig);
+        if (question) {
+          questions.push(question);
+        }
       }
     }
 
+    // Collect all answers (static + prompted)
+    let allAnswers = { ...staticAnswers };
+
     // Display appropriate header based on whether there are questions
     if (questions.length > 0) {
       CLIUtils.displayModuleConfigHeader(moduleName, moduleConfig.header, moduleConfig.subheader);
       console.log(); // Line break before questions
-      const answers = await inquirer.prompt(questions);
+      const promptedAnswers = await inquirer.prompt(questions);
 
-      // Store answers for cross-referencing
-      Object.assign(this.allAnswers, answers);
+      // Merge prompted answers with static answers
+      Object.assign(allAnswers, promptedAnswers);
+    }
+
+    // Store all answers for cross-referencing
+    Object.assign(this.allAnswers, allAnswers);
+
+    // Process all answers (both static and prompted)
+    if (Object.keys(allAnswers).length > 0) {
+      const answers = allAnswers;
 
       // Process answers and build result values
       for (const key of Object.keys(answers)) {

tools/cli/installers/lib/core/manifest-generator.js

@@ -581,6 +581,11 @@ class ManifestGenerator {
    */
   async writeWorkflowManifest(cfgDir) {
     const csvPath = path.join(cfgDir, 'workflow-manifest.csv');
+    const escapeCsv = (value) => `"${String(value ?? '').replaceAll('"', '""')}"`;
+    const parseCsvLine = (line) => {
+      const columns = line.match(/(".*?"|[^",\s]+)(?=\s*,|\s*$)/g) || [];
+      return columns.map((c) => c.replaceAll(/^"|"$/g, ''));
+    };
 
     // Read existing manifest to preserve entries
     const existingEntries = new Map();
@@ -592,18 +597,21 @@ class ManifestGenerator {
       for (let i = 1; i < lines.length; i++) {
         const line = lines[i];
         if (line) {
-          // Parse CSV (simple parsing assuming no commas in quoted fields)
-          const parts = line.split('","');
+          const parts = parseCsvLine(line);
           if (parts.length >= 4) {
-            const name = parts[0].replace(/^"/, '');
-            const module = parts[2];
-            existingEntries.set(`${module}:${name}`, line);
+            const [name, description, module, workflowPath] = parts;
+            existingEntries.set(`${module}:${name}`, {
+              name,
+              description,
+              module,
+              path: workflowPath,
+            });
           }
         }
       }
     }
 
-    // Create CSV header - removed standalone column as ALL workflows now generate commands
+    // Create CSV header - standalone column removed, everything is canonicalized to 4 columns
     let csv = 'name,description,module,path\n';
 
     // Combine existing and new workflows
@@ -617,12 +625,18 @@ class ManifestGenerator {
     // Add/update new workflows
     for (const workflow of this.workflows) {
       const key = `${workflow.module}:${workflow.name}`;
-      allWorkflows.set(key, `"${workflow.name}","${workflow.description}","${workflow.module}","${workflow.path}"`);
+      allWorkflows.set(key, {
+        name: workflow.name,
+        description: workflow.description,
+        module: workflow.module,
+        path: workflow.path,
+      });
     }
 
     // Write all workflows
     for (const [, value] of allWorkflows) {
-      csv += value + '\n';
+      const row = [escapeCsv(value.name), escapeCsv(value.description), escapeCsv(value.module), escapeCsv(value.path)].join(',');
+      csv += row + '\n';
     }
 
     await fs.writeFile(csvPath, csv);