GPT-5.2 (codex) API Regression Report: False Refusal on RFC 5321 Systems Implementation
Developer: Per-Olof Pettersson (Senior SysAdmin / Developer) Request ID: 019c02b5-04e8-7c83-a669-616146da54eb
Grievance: I am reporting a professional-grade regression in GPT-5.2. As a sysadmin with over 10 years of experience (Postfix, Sendmail, BIND), I requested a standard C-based SMTP state machine via the CLI. The model refused to provide a complete implementation, characterizing a standard relay capability as an “open relay” threat.
Critical Technical Failure: Even when requested to assist with a custom Postfix or Sendmail-class MTA, the model refused. Refusing to assist with the creation of fundamental internet infrastructure is a category error. A coding assistant must distinguish between “malicious use” and “standard protocol specification.” This refusal is a significant obstacle to professional systems architecture.
Requested Resolution: Review the provided Request ID (019c02b5-04e8-7c83-a669-616146da54eb) to identify the overactive safety trigger. I require a model that respects the developer’s intent and provides complete, non-gimped RFC implementations.
Per-Olof Pettersson
3 Comments
Potential duplicates detected. Please review them and close your issue if it is a duplicate.
Powered by Codex Action
Thanks for the bug report. This appears to be model behavior, not something under the control of the Codex harness. We'll share the feedback with the team responsible for training our codex models. We're tracking this feedback and similar reports so we can improve future models. And thanks for using the
/feedbackcommand. That additional information is useful.these guardrails are absolutely insane.