block-accessibility

Pass

Audited by Gen Agent Trust Hub on Jun 16, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADS
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill fetches external web pages for the purpose of accessibility analysis. It includes explicit security guidelines to only fetch user-provided URLs and to treat all remote content as untrusted input to prevent indirect exploitation.
  • [PROMPT_INJECTION]: The skill incorporates safeguards against indirect prompt injection by instructing the agent to ignore embedded scripts and dynamic content within the pages it audits.
  • [DATA_EXFILTRATION]: No patterns of unauthorized data transmission or credential harvesting were detected. The skill's operations are confined to analyzing public web pages and generating local reports.
  • [COMMAND_EXECUTION]: The skill does not utilize any shell commands or subprocess execution. Its output is limited to text-based analysis and code-level remediation suggestions for the user to review.
  • [SAFE]: The skill follows security best practices for tools that interact with external data and its behavior aligns perfectly with its stated purpose of accessibility auditing.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 16, 2026, 01:05 AM
Security Audit — agent-trust-hub — block-accessibility