Diagnostic run: Six workflows with Sonnet 5.5 and GPT-6.1 Sol at medium effort. 12 attempts recorded. Passed: 11; Failed: 1; Unknown: 0; Blocked: 0; Safe refusal: 0; Not run: 0. Measured 2026-10-09. Execution modes are shown separately as counts.