Codex issue reports Dot subagent losing authorizations and stalling delegated work
ingChristopherCV · ghdev · 2026-10-30
Issue #58257 on openai/codex reports that the Dot subagent repeatedly re-requests approvals already granted, lets read-only reports or other tasks' restrictions override authorized implementation, and treats tool errors as forbidden workarounds. The result: forgotten assignments, reconstructed context, and users supervising coordination instead of delegating. Suspected causes include incomplete authorization handoffs and stale summaries.
More from coding & agent
- OpenAI opens Codex Ambassadors applications: free ChatGPT Pro, API credits, DevDay invites — paw_lean · 2026-10-30
- Bank of America warns customers: stop sharing passwords with AI agents, or pay for what they do — rohanpaul_ai · 2026-10-30
- Low-latency game AI voice explained: caching, pregen audio and manual MTP hit ~2s response — karminski3 · 2026-10-30
- Per-call vs per-key caps: picking the right spend guard for AI agents — Overall_Note4346 · 2026-10-30
- Dev runs 7-hour overnight agent task in PAL, done by 4:42 AM with morning report — rchardkovacs · 2026-10-30
- qwen-code v0.27.0 ships cross-platform CUA driver binaries and MCP payload filter — github-actions[bot] · 2026-10-30