next-cache-components-optimizer

Pass

Audited by Gen Agent Trust Hub on Sep 4, 2026

Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTIONCREDENTIALS_UNSAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses the agent-browser CLI to automate browser actions, including setting cookies, navigating via pushstate, and taking screenshots.\n- [DYNAMIC_EXECUTION]: Employs the agent-browser eval tool to execute arbitrary JavaScript within the browser context, such as hiding the developer portal during screenshots.\n- [INDIRECT_PROMPT_INJECTION]: The skill's logic depends on parsing JSON output from agent-browser react suspense and server logs via mcp get_logs. This ingestion of environment data without explicit sanitization or boundary markers could allow malformed data to influence the agent's actions.\n
  • Ingestion points: Diagnostic JSON data from React Suspense boundaries and Next.js development logs.\n
  • Boundary markers: None identified in the instructions.\n
  • Capability inventory: Includes code refactoring (file system writes), browser cookie manipulation, and dynamic script execution in the browser.\n
  • Sanitization: The instructions do not specify any validation or filtering of the ingested data.\n- [CREDENTIALS_UNSAFE]: The skill operates within a browser session that holds authentication state and includes specific warnings to avoid clearing all cookies to prevent session loss, highlighting the sensitive nature of the execution environment.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 4, 2026, 02:25 PM
Security Audit — agent-trust-hub — next-cache-components-optimizer