search

Fail

Audited by Gen Agent Trust Hub on May 7, 2026

Risk Level: CRITICALPROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill uses the Bash(curl) tool to interact with the official serpapi.com API. This behavior is consistent with the skill's stated purpose and author context.
  • [SAFE]: Sensitive data is handled correctly. The skill instructs users to provide the SerpApi API key via an environment variable rather than hardcoding credentials.
  • [SAFE]: Automated scanner alerts (EshopScam and HttpRequest-inf) for serpapi.com are evaluated as false positives. The URLs point to official documentation for a legitimate search service, and the presence of HTTP requests is expected for this functionality.
  • [PROMPT_INJECTION]: The skill processes untrusted search results from various third-party engines, which represents a surface for indirect prompt injection.
  • Ingestion points: Search results fetched from the SerpApi endpoint (SKILL.md, examples.md).
  • Boundary markers: Absent; there are no specific instructions or delimiters used to isolate external data from the agent's internal reasoning.
  • Capability inventory: The skill utilizes the Bash(curl) tool for external communication and the Read tool for accessing local configuration schemas (SKILL.md).
  • Sanitization: No explicit sanitization or filtering of the external content is described before the agent processes and summarizes the results.
Recommendations
  • CRITICAL: 1 infected file(s) detected - DO NOT USE
  • Contains 2 malicious URL(s) - DO NOT USE
Audit Metadata
Risk Level
CRITICAL
Analyzed
May 7, 2026, 03:54 AM
Security Audit — agent-trust-hub — search