publish-x-article

Pass

Audited by Gen Agent Trust Hub on Mar 8, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill processes user-supplied Markdown content which can serve as a vector for indirect prompt injection.
  • Ingestion points: Markdown files read from local paths or URLs via bash commands and script processing.
  • Boundary markers: No delimiters or instructions to ignore embedded agent commands are applied to the processed article content.
  • Capability inventory: Extensive browser automation (navigation, clicking, typing), local script execution, file system access, and system clipboard operations.
  • Sanitization: Uses regex-based conversion in scripts/parse_markdown.py which lacks robust sanitization to prevent the execution of embedded instructional text intended to influence the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 8, 2026, 08:41 AM