qa-project-context

Pass

Audited by Gen Agent Trust Hub on May 11, 2026

Risk Level: SAFEPROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
  • [PROMPT_INJECTION]: The skill creates a centralized metadata file (.agents/qa-project-context.md) designed to be read by all other skills in the collection. This introduces a surface for indirect prompt injection. Ingestion point: .agents/qa-project-context.md. Boundary markers: Absent. Capability inventory: Downstream skills use this file to prioritize automation and configure test suites. Sanitization: Not provided.
  • [DATA_EXFILTRATION]: The skill gathers and stores sensitive project metadata including production and staging URLs, technology stack details, and high-impact risk areas. While intended for context sharing, this architectural documentation maps the project internal infrastructure and attack surface.
  • [SAFE]: The skill uses standard interactive discovery questions and project file scanning to assist the user. It does not perform unauthorized network requests or execute untrusted remote code.
Audit Metadata
Risk Level
SAFE
Analyzed
May 11, 2026, 07:46 AM