Plimsoll ships self-hosted sandbox for AI-generated code
What happened
On October 12, Hacker News featured Plimsoll's self-hosted sandbox for running code generated by AI agents. It offers four isolation levels: process, container, kernel and vm, and reports the actual level used on every run. If the sandbox cannot meet the requested minimum isolation level, it refuses to execute. API calls are authorized through a broker outside the sandbox, and run records get hash verification, signing and replay. The post covers the isolation, API authorization and record verification mechanisms.
Written by AI from the coverage · updated 57 minutes ago
Coverage
Follow the reports to see the story from different sides.
- Hacker News front pageShow HN: Plimsoll, a self-hosted sandbox for code your AI agent writes
Plimsoll 是用于运行 AI 智能体生成代码的自托管沙箱，每次返回实际隔离等级，达不到请求设定的下限就拒绝执行。它支持 process、container、kernel、vm 四级隔离，通过沙箱外的 broker 授权调用 API，并提供运行记录哈希校验、签名与重放。
Heat over time
Not enough continuous observations to draw a trend yet.