New user asks the agent security questions many avoid: default privileges, prompt injection, banking on same machine
meksilainen · reddit · 2026-11-12
An AI newcomer raising a series of substantive agent-security concerns: what privileges and data access do agents have by default, can guardrails against prompt injection actually be trusted, and is it risky to run agents on the same machine used for cloud storage and banking? They also weigh local setups like Ollama + OpenClaw versus Anthropic/OpenAI-hosted options. No answers yet, but the questions capture widespread user anxiety about agent privilege isolation worth attention from practitioners.
More from coding & agent
- TerraPulse: a global LiDAR archaeology platform inspired by 48-agent jungle scan of 1,200 km² — RichardsonDx · 2026-11-12
- Mashvisor ships an MCP server for rental property analytics, with setup walkthrough — ahmadhashlamoun · 2026-11-12
- Andrew Ng releases 4 free AI tools on GitHub, including desktop coworker OpenWorker — alex_verem · 2026-11-12
- The fastest way to make $0 with AI: automate outreach before your first client — aryanXmahajan · 2026-11-12
- Slaunt MCP server gives Claude Code 99% memory recall while cutting tokens 31% — karatyma_dev · 2026-11-12
- Managed agents platforms are the new chat-with-a-PDF, says developer — intellectronica · 2026-11-12