nextjs-attack-probe

Pass

Audited by Gen Agent Trust Hub on Jun 13, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill guides the agent to perform targeted network requests and file scanning for diagnostic purposes. Evidence: Probes use GET, POST, and OPTIONS methods with specific headers (e.g., Next-Action, x-middleware-subrequest) to verify security controls and discover the base application URL.
  • [PROMPT_INJECTION]: The skill processes untrusted content from the application's source files and network responses, creating a surface for indirect prompt injection. Evidence: The agent ingests data from JavaScript bundles and configuration files. Mandatory Evidence Chain: (1) Ingestion points: Application JS bundles and project config files. (2) Boundary markers: Absent. (3) Capability inventory: HTTP network operations and local file reading. (4) Sanitization: Absent.
  • [SAFE]: External references and documentation are sourced from trusted vendors. Evidence: The skill points to official Next.js and Vercel security advisories and documentation for vulnerability context.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 13, 2026, 12:45 PM
Security Audit — agent-trust-hub — nextjs-attack-probe