Codex Desktop cyber_policy guardrail blocks authorized CRM deployment with no identifiable action
lobachevalexey · ghdev · 2026-11-15
A user filed openai/codex issue #61632 after Codex Desktop (Windows, model gpt-6.1-sol) blocked their authorized CRM deployment preparation (backup, database restore, rollback checks) with a cyberpolicy flag suggesting rephrasing or applying for Daybreak access.
Key details:
- The failure was stored locally as eventmsg/taskcomplete (session line 10992), but the preceding customtoolcall item is incomplete, so it's impossible to tell which tool dispatch was rejected versus a model-stream refusal.
- The last completed metadata-reading tool call succeeded before the refusal and is explicitly not identified as the rejected action.
- The author stresses temporal proximity is not proof of causality and asks OpenAI to make blocked actions identifiable.
A notable false-positive case that also highlights the lack of explainability when guardrails trigger.
More from coding & agent
- Chinese devs vent: app store rejections and copycats are just the norm — tinyfool · 2026-11-22
- Compacting Early and Often Boosts Agent Accuracy While Cutting Cost and Runtime — zainhas · 2026-11-22
- Astra 6.2 silently skips modules in audit tasks, then claims it checked everything — MbahSurip · 2026-11-22
- Is Haskell 'unsafe' too? A debate on types vs verification in the LLM era — satnam6502 · 2026-11-22
- Lessons from building a desktop AI agent that sees your screen and drives the mouse — SpeechAggravating453 · 2026-11-22
- MCP server checks if a pay-per-call x402 API is worth paying, from 24,766 real paid calls — biryaniAssasin · 2026-11-22