prismor
Runtime security for AI coding agents that observes, asks for approval on, or blocks risky tool calls before they run, covering secret leaks, prompt injection, and supply chain risks through a local dashboard. Works across agents such as Claude and Codex.
1 skill470
@prismorsec