Which agent actions need human approval? Reddit debates pre-call guardrail layers
GLStops · reddit · 2026-11-11
A Reddit thread asks how to decide which agent actions need human approval when agents read emails, tickets, and web documents before acting. The poster worries about agents being influenced by outside content (e.g. prompt injection) and is considering building one or multiple pre-call guardrail layers to vet actions, without sandboxing away full capabilities. He asks what people do today and whether a tool here is worth building — pointing at the real unsolved problem of permission tiering and action-approval design in agent engineering.
Related event: Reddit debates which AI agent actions require human approval(2 posts)→
More from coding & agent
- KAIST's WikiTrace: evidence-linked memory fixes agent "memory-induced degradation" — dair_ai · 2026-11-11
- Why inference speed will be AI products' biggest competitive edge in 1–2 years — itsOmSarraf_ · 2026-11-11
- What happened when I moved vector retrieval into Snowflake SQL instead of orchestrating RAG outside — dima806_dima · 2026-11-11
- OpenAI engineer runs Codex for 26 days straight hunting Atlantis, spawning "vibe archaeology" — Polymarket · 2026-11-11
- Edward Hughes: hand-built AI harnesses have a ~3-month shelf life — Machine Learning Street Talk · 2026-11-11
- He replaced a $5,000/month ghostwriter with 11 Claude Skills and earned $47,000 from one post — aryanXmahajan · 2026-11-11