blog-geo

Warn

Audited by Gen Agent Trust Hub on Apr 9, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill includes instructions to execute shell commands using python3 to run a script located at skills/blog-google/scripts/run.py.
  • [COMMAND_EXECUTION]: The execution logic incorporates variables such as <url> and <property> into the command line arguments. If these variables are sourced from untrusted data (like the blog post being audited) without sanitization, it could lead to command injection.
  • [REMOTE_CODE_EXECUTION]: The skill relies on the existence and integrity of an external script (skills/blog-google/scripts/run.py) which is not part of the skill's own package, creating a dependency on external executable code.
  • [PROMPT_INJECTION]: The "Audit Process" involves reading and extracting the full content of blog posts. This data ingestion occurs without explicit boundary markers or instructions for the agent to ignore potentially malicious content embedded in the text, creating a surface for indirect prompt injection.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Apr 9, 2026, 02:45 AM