seo-audit-full
Pass
Audited by Gen Agent Trust Hub on Jun 16, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it crawls untrusted external content from a user-provided URL and processes it using multiple subagents and scripts.
- Ingestion points: The skill fetches HTML from a user-provided
[url]inscripts/fetch_page.pyand performs a deep crawl of up to 500 pages. - Boundary markers: There are no explicit instructions or delimiters defined to prevent the agent from following malicious instructions embedded within the crawled HTML content (e.g., hidden tags or metadata).
- Capability inventory: The skill possesses significant capabilities, including executing local Python scripts (
scripts/google_auth.py,scripts/drift_history.py, etc.), writing report files to disk, and capturing screenshots. - Sanitization: The skill does not mention any sanitization or filtering of the content retrieved from external websites before passing it to subagents like
seo-contentorseo-sxo. - [COMMAND_EXECUTION]: The skill executes shell commands that incorporate user-provided input. Specifically, the instruction
python scripts/drift_history.py <url>passes the raw URL argument directly to a Python script execution, which could lead to command injection if the environment does not properly sanitize the URL string.
Audit Metadata