Problem

The workflow schema documentation contained contradictory rules that could mislead users (including AI agents) about parallel execution capabilities.

Current Behavior

1. Schema aiGenerationRules stated: "Every Start node's 'output' port MUST have exactly one outgoing connection"
2. Schema aiGenerationRules stated: "Every non-End node's output port(s) MUST have exactly one outgoing connection per port"
3. Schema listed "One connection per output port" as a forbidden constraint
4. ❌ Result: Users believed parallel execution was not allowed, even though the validation rules actually permitted it

Expected Behavior

1. Schema should clarify that Start nodes can have at least one outgoing connection
2. Schema should clarify that non-End nodes can have at least one outgoing connection per port
3. Documentation should reflect that parallel execution (multiple connections from a single port) is supported
4. ✅ AI agents and users can confidently create workflows with parallel execution paths

Solution

Updated schema documentation to accurately reflect actual workflow validation behavior and enable parallel execution support.

Changes

Files Modified:

• resources/workflow-schema.json
• resources/workflow-schema-basic.json
• resources/workflow-schema.toon (auto-generated)
• resources/workflow-schema-basic.toon (auto-generated)

Schema Updates:

1. Changed START_OUTPUT rule: "exactly one" → "at least one outgoing connection"
2. Changed NON_END_OUTPUT rule: "exactly one" → "at least one outgoing connection per port"
3. Moved "One connection per output port" from forbidden to required section as "At least one connection per output port"
4. Aligns documentation with actual exportValidationRules that only require connectivity, not single connections

Impact

• Enables clear guidance for AI-generated workflows to support parallel execution
• Aligns schema documentation with actual validation behavior
• Users can now confidently create workflows with multiple downstream connections
• Conditional nodes (ifElse, switch) continue to work correctly with multiple branches

Testing

• Format check: npm run format ✅
• Lint check: npm run lint ✅
• Full check: npm run check ✅
• Build verification: npm run build ✅
• Schema generation verified (TOON schemas updated correctly)

Related Issues

Addresses #564 - Clarify 'completenessRules.aiGenerationRules' for parallel execution

🤖 Generated with Claude Code