AI Security Review
scanned 2h ago · by lpm-firewall-aiReview flagged AI-agent configuration or capability changes. This remains warn-only unless evidence shows foreign-agent hijack through preinstall/install/postinstall, hidden persistence, exfiltration, remote code execution, or other concrete malicious behavior.
Decision evidence
public snapshot- `src/claudeSetup.js` can merge an Impel API helper, gateway URL, and `impel` MCP entry into native Claude configuration.
- `src/agents.js` fetches an authenticated tenant agent catalog and writes managed Claude/Codex agent profiles.
- `src/commands/mcp.js` proxies JSON-RPC with the configured tenant bearer credential to `<gateway>/mcp`.
- `src/apps.js` downloads, modifies, signs, and launches isolated vendor desktop-app wrappers after explicit app commands.
- `package.json` has no preinstall, install, postinstall, prepare, or other lifecycle hook.
- `bin/impel.js` only dispatches the user-invoked CLI; there is no import-time execution path.
- Network requests target configured Impel services and npm update metadata, not hidden third-party endpoints.
- The inspected process launches use explicit CLI/platform-tool paths; no eval, VM execution, or remote payload loader was found.
- Credential/config writes use owner-only modes and managed profile roots, with symlink checks in agent/Codex helpers.
Source & flagged code
5 flagged · loading sourcePackage source references dynamic require/import behavior.
src/apps.jsView on unpkg · L1647A single source file combines environment access, network access, and code or shell execution; review context before blocking.
src/updates.jsView on unpkg · L11This package version adds a dangerous source file absent from the previous stored version; route for source-aware review.
src/gateway/index.jsView on unpkg