AI agents can merge code to main unattended but can't send an email: a paradox of safety boundaries
themaxthule · reddit · 2026-08-01
The author shares a thought-provoking engineering practice: his AI coding agents can merge code to main branch unattended, but their permissions are so restricted they can't even send an email. He admits merging unreviewed code is more dangerous than sending an email, yet maintains this seemingly contradictory permission setup.
During an audit, he discovered his monitoring tool had blind spots: tagging rules based on branch prefixes (like claude/ or codex/) missed some agent operations because agents run under a normal account. The fix was an allowlist file, but the directory containing it was gitignored due to a private key, so the fix itself couldn't be version-controlled.
He poses an open question: what would have to be true for you to let one of your workflows write more autonomously than it does today?
More from coding & agent
- GitHub rewrote the 800k-line Copilot runtime in Rust using AI agents, cutting CPU and memory by over 90% — bibryam · 2026-10-16
- What's the hardest part of deploying AI in real businesses that nobody talks about — djgoel123 · 2026-10-16
- DeepMind turns coding agent Antigravity into a science tool, surfacing novel antimicrobial peptide in one session — pushmeet · 2026-10-16
- Cross-review coding workflow: Claude writes the plan, Codex audits it — theteknosaur · 2026-10-16
- NPCTalk makes GTA V NPCs conversable in ~1s, open-sourced for any game — ChocolateDull8971 · 2026-10-16
- Simon Willison: code mode is the future of tool calling, if it can be sandboxed — springrod · 2026-10-16