seo-performance

Pass

Audited by Gen Agent Trust Hub on May 16, 2026

Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes shell commands like mkdir, grep, and echo to manage a local .seo-cache directory and update the .gitignore file as part of its audit workflow.\n- [DATA_EXFILTRATION]: The skill performs network operations to fetch performance metrics from user-provided URLs and the Google PageSpeed Insights API. Auditing external sites is the core function of the skill.\n- [REMOTE_CODE_EXECUTION]: The skill instructs the agent to execute a local Python script, scripts/analyze_performance.py, to conduct the audit analysis.\n- [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface because it ingests and processes untrusted content from the web.\n
  • Ingestion points: Performance metrics and site data fetched from arbitrary external URLs and API responses.\n
  • Boundary markers: No specific delimiters or protective instructions are provided to isolate the ingested data from the agent's core logic.\n
  • Capability inventory: The skill is capable of executing local scripts, running shell commands, and writing files to the disk.\n
  • Sanitization: No data validation or sanitization steps are documented for the content retrieved from external sources.
Audit Metadata
Risk Level
SAFE
Analyzed
May 16, 2026, 08:38 AM
Security Audit — agent-trust-hub — seo-performance