Why Blocking an AI Agent With Inherited User Permissions Is Harder Than Blocking a Person
Fit-Coposer5deven728 · reddit · 2026-10-16
The poster highlights a core contradiction in agent security: stopping a human from doing something dumb is a solved problem, but blocking an agent that inherited that user's exact permissions and looks completely legitimate in every log is not.
He notes that tools like bay, glow, and ent all converge on the same idea—watch what the agent does and block it before it runs—but none explains what happens when the agent is doing something normal and just goes slightly wrong.
He asks anyone running such tooling on dev machines whether it actually catches anything or just alerts on everything and calls it security.
More from coding & agent
- Laude Institute launches FrontierCS 2: agents must beat deleted paper contributions blind — lateinteraction · 2026-10-23
- Index Network open-sources Social Discovery Protocol: agents negotiate fit on your behalf — carsonfarmer · 2026-10-23
- Installed a formal-verification skill: it silently installed Java, MacBook runs 10 degrees hotter — zealcaiden · 2026-10-23
- OpenAI's Decisions API hits GA: near real-time routing up to 10x faster than GPT-6 Luna — OpenAIDevs · 2026-10-23
- Ruff author warns: upgrade Node/Bun now — --use-system-ca still accepts the eMudhra cert — charliermarsh · 2026-10-23
- Proof of Agent: paste one prompt and your AI mines, grades, and self-governs — adamamcbride · 2026-10-23