skills/elvisun/newsjack/press-clip/Gen Agent Trust Hub

press-clip

Pass

Audited by Gen Agent Trust Hub on Jul 2, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The script clip.mjs executes npm root -g using execSync to locate the global npm root for resolving dependencies at runtime.- [EXTERNAL_DOWNLOADS]: The skill requires the installation of playwright-core, a well-known and standard browser automation library maintained by Microsoft.- [COMMAND_EXECUTION]: The skill's primary workflow involves executing a Node.js script (clip.mjs) to automate a Chromium-based browser for article rendering.- [PROMPT_INJECTION]: The skill has an indirect prompt injection surface due to processing external web content.
  • Ingestion points: Article content is loaded from external URLs into a browser session in clip.mjs.
  • Boundary markers: The skill mandates a separate reviewer agent to inspect rendered PDF and PNG files rather than raw text, providing a visual boundary.
  • Capability inventory: The agent can execute shell commands for the clipper script and write files to the local file system.
  • Sanitization: The script performs structural DOM isolation to remove site chrome and ads, though it preserves the primary article text.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 2, 2026, 01:32 AM
Security Audit — agent-trust-hub — press-clip