skills/lijigang/ljg-skills/ljg-card/Gen Agent Trust Hub

ljg-card

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill requires the Playwright library and the Chromium browser, which are fetched from official package registries using standard bun/npm commands as part of the tool's environment setup.
  • [COMMAND_EXECUTION]: The skill executes local TypeScript utilities (capture.ts, prepare-whiteboard-source.ts, and verify-full-text.ts) via the Bun runtime. These scripts perform the core logic of source preparation, content verification, and HTML-to-PNG rendering.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from URLs, files, and user text. It mitigates injection risks through content verification scripts that ensure output integrity at the character level. Ingestion points: web reading tools, pasted text, and local files mentioned in SKILL.md. Boundary markers: HTML safe escaping and logic ledger matching. Capability inventory: subprocess spawning via Bun and browser automation via Playwright. Sanitization: character entity escaping and source-ledger consistency checks.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 02:40 AM
Security Audit — agent-trust-hub — ljg-card