seo-geo-skills
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [NO_CODE] (SAFE): The skill is entirely composed of documentation and prompt templates in Markdown format. No executable scripts (.py, .js, .sh), binaries, or configuration files that could trigger code execution were found.
- [Indirect Prompt Injection] (LOW): The skill is designed to have the agent ingest and analyze content from untrusted external sources (competitor URLs and user-provided domains). This creates a surface for indirect prompt injection if those websites contain hidden malicious instructions.
- Ingestion points: [competitor URL], [URL], and [domain] parameters used in research and audit prompts (SKILL.md).
- Boundary markers: Absent; the prompts do not include instructions for the agent to ignore embedded commands within the fetched web content.
- Capability inventory: No active code capabilities (subprocesses/network calls) are defined within the skill itself; it relies on the agent's native browsing abilities.
- Sanitization: Absent; there is no mention of filtering or sanitizing the retrieved HTML or text before processing.
Audit Metadata