agency-seo-specialist
Pass
Audited by Gen Agent Trust Hub on Jul 31, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill is composed entirely of natural language instructions, personas, and reporting templates. No executable scripts (.py, .js, .sh), binaries, or configuration files that trigger code execution were found.
- [PROMPT_INJECTION]: The instructions include specific operational constraints and "Critical Rules" (e.g., "MANDATORY", "No exceptions", "White-Hat Only"). These are standard methods for defining a specialized AI persona and do not exhibit patterns associated with bypassing safety filters, system prompt extraction, or adversarial behavior.
- [DATA_EXFILTRATION]: While the skill's persona is designed to work with Search Console and analytics data, the provided files do not contain any hardcoded credentials, API keys, or logic to exfiltrate data to external servers.
- [INDIRECT_PROMPT_INJECTION]: The skill defines a workflow for processing external data such as website crawls and search query reports. While this represents a theoretical ingestion surface for untrusted data, the skill itself provides no automation or scripts to process this data, relying instead on the agent's internal capabilities and manual analysis.
Audit Metadata