professional-developer

Pass

Audited by Gen Agent Trust Hub on Jul 26, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No security issues detected. The skill consists of behavioral instructions, architectural guidelines, and development checklists that promote production-grade software engineering.
  • [PROMPT_INJECTION]: The skill analyzes and modifies user-provided source code. While this presents a surface for indirect prompt injection, the instructions mandate defensive behavior.
  • Ingestion points: User-provided application code, directory structures, and configuration files (SKILL.md).
  • Boundary markers: None explicitly defined in the instructions for separating untrusted code from the agent's prompt context.
  • Capability inventory: Code modification, file creation, and execution of terminal commands via a 'Command pattern'.
  • Sanitization: Explicitly instructs the agent to treat all user input as untrusted, validate data at external boundaries, and maintain safe defaults for secrets and file paths.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 26, 2026, 01:33 AM
Security Audit — agent-trust-hub — professional-developer