local-seo
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill analyzes external website content provided by users (via URLs), which creates an attack surface for instructions embedded in the target site (e.g., in HTML comments or JSON-LD) to manipulate the agent. \n
- Ingestion points: User-provided business URLs defined in the argument-hint field of SKILL.md.\n
- Boundary markers: The skill instructions do not define delimiters or specific safety instructions to isolate external content from the agent's logic.\n
- Capability inventory: The agent performs analysis, diagnoses ranking factors, and generates fix plans based on the ingested data.\n
- Sanitization: No mention of content sanitization, filtering, or validation is present in the skill instructions.\n- [DYNAMIC_EXECUTION]: The skill uses a workflow redirection pattern that instructs the agent to read and follow instructions from a separate file located at ../../seo/local-seo/SKILL.md. This constitutes dynamic loading of instructions from a location outside the immediate skill directory.
Audit Metadata