web-development

Pass

Audited by Gen Agent Trust Hub on Sep 27, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill references the CloudBase Web SDK from the vendor's official CDN at https://static.cloudbase.net/cloudbase-js-sdk/latest/cloudbase.full.js for use in static web pages.
  • [COMMAND_EXECUTION]: The skill requires the execution of standard build and linting commands, including npm run build, vite build, tsc --noEmit, and eslint, as part of its mandatory self-verification process.
  • [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to ingest untrusted data from web pages via a browser tool, which could be used for indirect prompt injection attacks.
  • Ingestion points: The agent-browser tool is used to access local dev servers or external sites to reproduce bugs and verify features (specified in browser-testing.md).
  • Boundary markers: None identified; no specific guidance is given to distinguish between instructions and data when interacting with browser content.
  • Capability inventory: The skill has capabilities for file system modification, command execution, and cloud platform management using tools like manageHosting and manageFunctions (specified in SKILL.md and frameworks.md).
  • Sanitization: No validation or sanitization requirements are mentioned for the content retrieved through the browser validation flow.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 27, 2026, 09:21 AM
Security Audit — agent-trust-hub — web-development