MCP/CLI Parity Triage Use this when a bug report says "works in CLI, fails in MCP" or the reverse. Workflow 1. Start from the issue payload and identify the exact tool/action pair. 2. Run the narrowest existing MCP regression bucket for that payload. 3. Run the matching CLI parity bucket sequentially (not in parallel with another build/test run). 4. Interpret the split: - MCP fails, CLI passes: inspect MCP schema, hand-written tool routing, service forwarding, and response envelope handling. - CLI fails, MCP passes: inspect CLI command wiring, daemon/service transport, and exit-code/error ser…