AI Security Review
scanned 8d ago · by lpm-firewall-aiNo confirmed malicious attack surface. Observed network and storage primitives support a whiteboard/collaboration library and require runtime host/user configuration or user actions.
Decision evidence
public snapshot- package.json has no install/preinstall/postinstall/prepare lifecycle scripts; published entrypoints are dist/index.cjs.js and dist/index.js.
- dist/index.cjs.js and loader/index.cjs.js only delegate to bundled CJS files; no install-time execution or shell execution found.
- Trojan Source check over dist/esm/schema.constants-CH2FWTVr.js and dist/cjs/schema.constants-DItwOm2z.js found no bidi/invisible control characters.
- Network code is package-aligned: optional collaboration providers default to ws://localhost:1234 and asset providers use host-supplied URLs.
- License code verifies an offline Ed25519 public key and schedules local revalidation; it does not exfiltrate keys or credentials.
- Searches found no child_process, credential harvesting, persistence, destructive file writes, or AI-agent control-surface mutation.
Source & flagged code
3 flagged · loading sourcePackage source references dynamic require/import behavior.
dist/index.cjs.jsView on unpkg · L1Source contains bidi control or invisible Unicode characters associated with Trojan Source attacks.
dist/esm/schema.constants-CH2FWTVr.jsView on unpkg · L2602A package entrypoint or install-time lifecycle script reaches a source file with blocking dangerous behavior.
dist/cjs/schema.constants-DItwOm2z.jsView on unpkg