debug-automation-tests

Pass

Audited by Gen Agent Trust Hub on Jun 19, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill defines a standard debugging workflow using five purpose-built MCP tools to fetch test data and logs from the LambdaTest platform.
  • [PROMPT_INJECTION]: The skill was evaluated for indirect prompt injection risks.
  • Ingestion points: Automation test metadata and logs (command, network, console, Selenium) are fetched into the context via MCP tools.
  • Boundary markers: The instructions do not specify the use of delimiters for the retrieved log data.
  • Capability inventory: The skill does not incorporate tools that allow for command execution, file system writes, or arbitrary network access.
  • Sanitization: There is no evidence of log content sanitization.
  • Assessment: Due to the absence of dangerous tools that could be manipulated by data, the surface for indirect prompt injection is considered safe.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 19, 2026, 06:51 AM
Security Audit — agent-trust-hub — debug-automation-tests