tk-browser-verify

Pass

Audited by Gen Agent Trust Hub on Sep 8, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No malicious patterns or security risks were detected. The skill follows best practices for tool execution and data handling.
  • [INDIRECT_PROMPT_INJECTION]: The skill manages a significant attack surface by ingesting web content, logs, and command outputs. It mitigates this via a mandatory 'Retrieved Evidence Boundary' instruction and explicit redaction of sensitive network data. Ingestion points include web/file content and command output; capabilities are limited to browser automation and run-owned background servers.
  • [COMMAND_EXECUTION]: Local development servers (e.g., react-scripts) are managed as run-owned background processes with readiness checks, avoiding arbitrary command injection by sticking to verified repository scripts.
  • [CREDENTIALS_UNSAFE]: Authentication is handled through a secure local file mechanism (0600 permissions) and a temporary loopback server for injection, preventing secrets from being exposed in conversation logs or command-line history.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 8, 2026, 10:19 AM
Security Audit — agent-trust-hub — tk-browser-verify