agent-guidance
Pass
Audited by Gen Agent Trust Hub on Oct 20, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: No shell commands or script execution patterns were found in the skill or its references.
- [DATA_EXFILTRATION]: No network operations or sensitive file access patterns were detected. External links are limited to vendor-owned repositories and well-known academic or professional resources (e.g., Anthropic, NN/g, arXiv).
- [REMOTE_CODE_EXECUTION]: The skill does not download or execute any external scripts or packages.
- [INDIRECT_PROMPT_INJECTION]: The skill provides instructions for the agent to review and design guidance based on user-provided task descriptions. While this involves processing external data, the skill does not have capabilities (such as file writes, network operations, or subprocess execution) that would make it vulnerable to exploitation via indirect injection.
- [OBFUSCATION]: No encoded content, hidden characters, or homoglyphs were detected.
Audit Metadata