skills/yonatangross/orchestkit/dev/Gen Agent Trust Hub

dev

Pass

Audited by Gen Agent Trust Hub on May 4, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDATA_EXFILTRATION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes various system and development commands through its scripts. It runs portless, emulate, agent-browser, and the project's own dev scripts (via pnpm, npm, yarn, or bun).
  • Evidence: scripts/boot.sh constructs and executes commands like portless "${slug}" "${pkg_mgr}" run dev.
  • Mitigation: The script implements sanitization for the branch-based slug (using tr and cut) to prevent command injection via malicious branch names.
  • [EXTERNAL_DOWNLOADS]: The skill uses npx --yes wait-on to poll the dev server's availability. npx downloads the package from the npm registry at runtime if not locally cached.
  • Evidence: Found in scripts/boot.sh at line 231: npx --yes wait-on --httpTimeout 30000 --tlsCheck false "${base_url}".
  • [DATA_EXFILTRATION]: The --funnel and --live flags utilize portless and tailscale to expose the local development server to the public internet via a tunnel.
  • Evidence: scripts/boot.sh implements these flags and provides a warning: ⚠ funnel mode is PUBLIC on the internet. Anyone with the URL can reach the dev server.
  • Context: This is an intentional feature for sharing previews but represents a significant security surface if sensitive data or unauthenticated endpoints are exposed.
Audit Metadata
Risk Level
SAFE
Analyzed
May 4, 2026, 05:09 PM