skills/richardbray/skills/questions/Gen Agent Trust Hub

questions

Pass

Audited by Gen Agent Trust Hub on Jul 10, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No malicious patterns or security risks were identified. The skill is entirely text-based and contains no executable scripts, shell commands, or external dependencies.
  • [PROMPT_INJECTION]: The instructions establish conversational triggers and guidance for agent behavior without attempting to override system safety filters or engage in jailbreak behavior.
  • [INDIRECT_PROMPT_INJECTION]:
  • Ingestion points: User-provided messages and requests are analyzed for ambiguity in SKILL.md.
  • Boundary markers: None provided in the prompt logic.
  • Capability inventory: No tools, file system access, or network capabilities are requested or used by the skill.
  • Sanitization: None; however, the absence of actionable capabilities prevents the execution of embedded instructions.
  • [EXTERNAL_DOWNLOADS]: The README.md file contains an installation command referencing the author's GitHub repository. This is an installation instruction for the user and does not represent a runtime download or remote code execution risk by the skill itself.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 10, 2026, 08:53 AM
Security Audit — agent-trust-hub — questions