keyword-clustering
Pass
Audited by Gen Agent Trust Hub on Aug 20, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [PROMPT_INJECTION]: No malicious instructions, behavior overrides, or safety filter bypasses were detected.
- [DATA_EXFILTRATION]: No access to sensitive local files or unauthorized network activity was identified. Data operations are confined to defined SEO research tools.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted search result data, creating a potential but low-risk surface for indirect prompt injection. 1. Ingestion points: get_serp_results, get_local_serp_results, get_search_console_performance, research_keywords, and get_ranked_keywords (SKILL.md). 2. Boundary markers: Absent. 3. Capability inventory: save_keywords (SKILL.md). 4. Sanitization: Absent.
- [OBFUSCATION]: No obfuscated content or hidden URLs were found.
- [REMOTE_CODE_EXECUTION]: The skill does not include any external dependencies or commands for executing remote code.
Audit Metadata