.claude/workflows/{category}/{workflow-name}.md

1. {Phase Name} - {Description} ...

• {agent-name}

  - {Role in workflow}

**Verification:**

```bash
grep "{workflow-name}" .claude/CLAUDE.md || echo "ERROR: CLAUDE.md SECTION 8.6 NOT UPDATED!"

1. User requests complex orchestration requiring multiple agents
2. Repeatable multi-phase process needs standardization
3. Router identifies workflow gap for common task pattern
4. Existing workflow needs enhancement or customization

.claude/docs/@WORKFLOW_AGENT_MAP.md

• Section 1: Full workflow-agent matrix
• Section 2: Archetype workflow sets (Router/Orchestrator, Implementer, Reviewer, Documenter, Researcher, Domain)

## Related Workflows

1. Check if workflow already exists:
   bash

   test -f .claude/workflows/<category>/<workflow-name>.md && echo "EXISTS" || echo "NEW"

2. If workflow EXISTS:
   • DO NOT proceed with creation
   • Invoke artifact-updater workflow instead:
     javascript

     Skill({
       skill: 'artifact-updater',
       args: {
         name: '<workflow-name>',
         changes: '<description of requested changes>',
         justification: 'Update requested via workflow-creator',
       },
     });

   • Return updater result and STOP
3. If workflow is NEW:
   • Continue with Step 0.5 below

1. Use

   companion-check.cjs

   from

   .claude/lib/creators/companion-check.cjs

2. Call

   checkCompanions("workflow", "{workflow-name}")

   to identify companion artifacts
3. Review the companion checklist — note which required/recommended companions are missing
4. Plan to create or verify missing companions after this artifact is complete
5. Include companion findings in post-creation integration notes

• {value_1}: {Description}
• {value_2}: {Description}

1. Read your agent definition: .claude/agents/{category}/{agent}.md
2. Invoke skills: Skill({ skill: "{skill-name}" })
3. {Additional instructions}
4. Save output to: .claude/context/{output-path}

• {Context item 1}
• {Context item 2}

1. Read .claude/context/memory/learnings.md first
2. Record decisions to .claude/context/memory/decisions.md `, });

undefined

1. {Recovery step}
2. Restart Phase 1

1. {Recovery step}
2. Re-run reviews

• {Criterion 1}
• {Criterion 2}
• All agents completed their phases
• Memory files updated

• {Parameter 1}
• {Parameter 2}

• model field MUST be base name only:

  haiku

  ,

  sonnet

  , or

  opus

• DO NOT use dated versions like

  claude-opus-4-5-20251101

• Choose model based on task complexity:

  • haiku

    : Quick validation, simple tasks

  • sonnet

    : Standard work (default)

  • opus

    : Complex reasoning, architecture

// CORRECT
Task({
  task_id: 'task-5',
  model: 'sonnet',
  // ...
});

// WRONG - causes validation failures
Task({
  task_id: 'task-6',
  model: 'claude-sonnet-4-20250514',
  // ...
});

• Use standard tools: Read, Write, Edit, Bash, TaskUpdate, TaskList, etc.
• DO NOT include MCP tools (mcp__*) unless whitelisted in routing-table.cjs
• MCP tools cause router enforcement failures

allowed_tools: [
  'Read',
  'Write',
  'Edit',
  'Bash',
  'TaskUpdate',
  'TaskList',
  'TaskCreate',
  'TaskGet',
  'Skill',
];

• mcp__Exa__*

  - Only for evolution-orchestrator

• mcp__github__*

  - Only for specific GitHub operations
• Other MCP tools - Must be whitelisted in routing-table.cjs

[ ] Model field uses base name (haiku/sonnet/opus)
[ ] allowed_tools contains only standard tools OR explicitly whitelisted MCP tools
[ ] No dated model versions (e.g., claude-opus-4-5-20251101)

1. Find Section 8.6 ("ENTERPRISE WORKFLOWS")
2. Generate workflow entry in this format:

undefined

.claude/workflows/{category}/{workflow-name}.md

1. {Phase 1 Name} - {Brief description}
2. {Phase 2 Name} - {Brief description} ...

• {Option category}:

  {values}

• {agent-name}

  - {Role description}

3. **Insert in appropriate location** (alphabetical or at end of section)
4. **Verify with**:

```bash
grep "{workflow-name}" .claude/CLAUDE.md || echo "ERROR: CLAUDE.md NOT UPDATED - BLOCKING!"

1. Create/Update Workflow Registry Entry in appropriate location:
   If registry doesn't exist, create

   .claude/context/artifacts/catalogs/workflow-registry.json

   :
   json

   {
     "workflows": [
       {
         "name": "{workflow-name}",
         "id": "{workflow-name}",
         "description": "{Brief description from workflow}",
         "category": "{enterprise|core|operations}",
         "version": "1.0.0",
         "triggerConditions": ["{Condition 1}", "{Condition 2}"],
         "participatingAgents": ["{agent-1}", "{agent-2}"],
         "phases": [
           { "name": "{Phase 1 Name}", "description": "{Brief description}" },
           { "name": "{Phase 2 Name}", "description": "{Brief description}" }
         ],
         "filePath": ".claude/workflows/{category}/{workflow-name}.md",
         "orchestratorReferences": []
       }
     ]
   }

2. Register with Appropriate Orchestrator:
   Update the orchestrator that should use this workflow:
   • master-orchestrator.md: Meta-workflows coordinating other workflows
   • evolution-orchestrator.md: EVOLVE phase workflows (research, creation, validation)
   • party-orchestrator.md: Consensus/voting workflows
   • swarm-coordinator.md: Coordination workflows
   Add to orchestrator's workflow references:
   markdown

   ## Available Workflows
   
   - `{workflow-name}` (.claude/workflows/{category}/{workflow-name}.md)

3. Document in

   .claude/docs/WORKFLOW_CATALOG.md

   :
   If catalog doesn't exist, create it. Add entry:
   markdown

   ### {Workflow Title}
   
   **Path:** `.claude/workflows/{category}/{workflow-name}.md`
   
   **When to use:** {Trigger conditions - specific scenarios}
   
   **Participating Agents:**
   
   - {agent-1}: {role in workflow}
   - {agent-2}: {role in workflow}
   
   **Phases:**
   
   1. **{Phase 1 Name}** - {Description}
   2. **{Phase 2 Name}** - {Description}
   
   **Orchestrator Integration:**
   
   - Invoked by: {orchestrator-name}
   - Trigger: {How orchestrator discovers/triggers this}

4. Verify Workflow Discoverability:
   Test that the workflow can be discovered:
   bash

   # Check workflow registry exists and is valid JSON
   node -e "console.log(JSON.stringify(require('./.claude/context/artifacts/catalogs/workflow-registry.json'), null, 2))"
   
   # Verify orchestrator references it
   grep "{workflow-name}" .claude/agents/orchestrators/*-orchestrator.md

5. Update Memory:
   Append to

   .claude/context/memory/learnings.md

   :
   markdown

   ## Workflow: {workflow-name}
   
   - **Purpose:** {Workflow purpose}
   - **Orchestrator:** {Which orchestrator uses it}
   - **Key Pattern:** {Key pattern discovered}
   - **Integration Notes:** {Any special integration considerations}

• Orchestrators cannot discover available workflows
• Workflows cannot be dynamically selected for complex tasks
• System loses visibility into workflow network
• "Invisible artifact" pattern emerges

TaskUpdate({ status: "completed" })

1. Run the 10-item integration checklist:
   bash

   node .claude/tools/cli/validate-integration.cjs .claude/workflows/<category>/<workflow-name>.md

2. Verify exit code is 0 (all checks passed)
3. If exit code is 1 (one or more checks failed):
   • Read the error output for specific failures
   • Fix each failure:
     • Missing CLAUDE.md entry -> Add to Section 8.6
     • Missing workflow registry -> Create registry entry (Step 8)
     • Missing orchestrator reference -> Add to orchestrator definition
     • Missing memory update -> Update learnings.md
   • Re-run validation until exit code is 0
4. Only proceed when validation passes

.claude/workflows/core/post-creation-validation.md

.claude/workflows/core/router-decision.md

• How the Router discovers and invokes this skill's artifacts

.claude/workflows/core/skill-lifecycle.md

• Discovery, creation, update, deprecation phases
• Version management and registry updates
• CLAUDE.md integration requirements

.claude/workflows/core/external-integration.md

• Safe integration of external artifacts
• Security review and validation phases

1. Identify phases: Discovery -> Architecture -> Implementation -> Testing -> Deployment
2. Assign agents: planner -> architect + security-architect -> developer -> qa -> devops
3. Define handoffs: Each phase outputs to

   .claude/context/plans/feature-<name>/

4. Reference: See

   .claude/workflows/enterprise/feature-development-workflow.md

1. Identify phases: Triage -> Investigation -> Resolution -> Post-mortem
2. Assign agents: incident-responder -> devops-troubleshooter + developer -> devops -> planner
3. Define handoffs: Each phase outputs to

   .claude/context/incidents/

4. Create:

   .claude/workflows/operations/incident-response.md

1. Identify phases: Code Analysis -> Component Synthesis -> Container Mapping -> Context Documentation
2. Assign agents: c4-code -> c4-component -> c4-container -> c4-context
3. Define handoffs: Bottom-up, each level builds on previous documentation
4. Reference: See

   .claude/workflows/enterprise/c4-architecture-workflow.md

.claude/templates/workflows/workflow-template.md

1. YAML Frontmatter: name, description, version, agents, tags
2. Overview: Purpose and when to use
3. Configuration Options: Customizable parameters
4. Phases: Sequential or parallel execution blocks
5. Task Spawns: Actual Task() calls with full prompts
6. Error Recovery: What to do when phases fail
7. Success Criteria: How to verify workflow completed successfully
8. Usage Example: How Router should invoke the workflow

• Workflows:

  .claude/workflows/{category}/

  (core, enterprise, operations, creators)
• Tests:

  tests/

  (NOT in .claude/)
• Related agents:

  .claude/agents/{category}/

• Related skills:

  .claude/skills/{name}/

• Related templates:

  .claude/templates/

• Reference files use @notation: @ENTERPRISE_WORKFLOWS.md, @AGENT_ROUTING_TABLE.md
• Located in:

  .claude/docs/@*.md

• See: CLAUDE.md Section 8.6 (ENTERPRISE WORKFLOWS reference)

• Workflow spawn templates must include:

  cd "$PROJECT_ROOT" || exit 1

  for background tasks
• Environment variables control validators (block/warn/off mode)
• See: .claude/docs/SHELL-SECURITY-GUIDE.md
• Apply to: agent spawn prompts, task execution examples

• ADR-075: Router Config-Aware Model Selection
• ADR-076: File Placement Architecture Redesign
• ADR-077: Shell Command Security Architecture

.claude/workflows/<category>/

• core/

  - Essential workflows (router-decision, skill-lifecycle)

• enterprise/

  - Complex multi-agent flows (feature-development)

• operations/

  - Operational procedures (incident-response)

• creators/

  - Creator workflow YAML files

• File Placement: See

  .claude/docs/FILE_PLACEMENT_RULES.md

• Developer Workflow: See

  .claude/docs/DEVELOPER_WORKFLOW.md

• Artifact Naming: See

  .claude/docs/ARTIFACT_NAMING.md

cat .claude/context/memory/learnings.md

• Previously created workflows
• Workflow patterns that worked well
• Issues with existing workflows

• New workflow created -> Append to

  .claude/context/memory/learnings.md

• Workflow issue found -> Append to

  .claude/context/memory/issues.md

• Architecture decision -> Append to

  .claude/context/memory/decisions.md

ASSUME INTERRUPTION: Your context may reset. If it's not in memory, it didn't happen.

1. Call

   runIntegrationChecklist(artifactType, artifactPath)

   from

   .claude/lib/creators/creator-commons.cjs

2. Call

   queueCrossCreatorReview(artifactType, artifactPath)

   from

   .claude/lib/creators/creator-commons.cjs

3. Review the impact report — address all

   mustHave

   items before marking task complete
4. Log any

   shouldHave

   items as follow-up tasks

• Workflow added to @WORKFLOW_AGENT_MAP.md
• Workflow referenced in CLAUDE.md (if enterprise workflow)
• Workflow assigned to at least one agent
• Workflow catalog entry added (if applicable)