seo-content-service-page
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute local Python scripts using shell commands (e.g.,
python3 skills/seo-geo-audit/scripts/seo_audit.py [URL]). User-provided URLs are directly interpolated into these commands. This pattern presents a risk of command injection if the agent does not sanitize the input, allowing an attacker to append malicious shell operators (e.g.,;,&,|) to the URL string. - [INDIRECT_PROMPT_INJECTION]: The skill possesses a significant attack surface for indirect prompt injection by processing external and local untrusted data.
- Ingestion points: The skill reads and audits content from user-provided URLs and local knowledge bases, such as Obsidian vaults (markdown files).
- Boundary markers: The instructions lack defined delimiters or specific directives for the agent to ignore instructions embedded within the data being analyzed.
- Capability inventory: The skill allows for the execution of local scripts via the shell and the generation of content based on the analyzed data.
- Sanitization: No mechanisms for sanitizing, validating, or escaping the content from external pages or local notes are described in the instructions.
Audit Metadata