test-it

Pass

Audited by Gen Agent Trust Hub on Oct 23, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes external code, test suite outputs, and evaluation datasets (Regime C), which creates a surface for indirect prompt injection where malicious instructions could be embedded in data processed by the agent.
  • Ingestion points: Repository files, test output logs, and external evaluation datasets (SKILL.md).
  • Boundary markers: The skill does not explicitly define delimiters for untrusted data, though it mandates a structured workflow and output format.
  • Capability inventory: The skill can execute shell commands (test runners), read and write to the file system, and spawn subagents (SKILL.md).
  • Sanitization: The instructions include a specific guardrail to redact secrets in fixtures and snapshots, mitigating credential exposure risks (SKILL.md).
  • [EXTERNAL_DOWNLOADS]: The skill references integration with well-known services and frameworks such as Langfuse, OpenTelemetry, and Playwright for evaluation and testing. These are industry-standard tools for the described workflow and are considered safe external references.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 23, 2026, 01:31 AM