occlusion-trace-meta-proof
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: LOWNO_CODE
Full Analysis
- NO_CODE (SAFE): The skill lacks any executable code, shell scripts, or external dependencies. It relies entirely on natural language instructions for the AI agent.
- PROMPT_INJECTION (SAFE): No malicious instructions designed to bypass safety guardrails or hijack the agent's persona were identified.
- DATA_EXFILTRATION (SAFE): The skill does not contain any commands to access sensitive files or transmit data to external servers.
- EXTERNAL_DOWNLOADS (SAFE): No remote resources or third-party packages are referenced or downloaded.
Audit Metadata