tool-calling
Pass
Audited by Gen Agent Trust Hub on Aug 22, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides documentation and Python usage examples for the 'liter-llm' library, which is a resource associated with the vendor 'xberg-io'.
- [SAFE]: The provided code examples demonstrate secure credential management by utilizing environment variables (os.environ['OPENAI_API_KEY']) instead of hardcoding sensitive keys.
- [PROMPT_INJECTION]: The skill documents a workflow for processing tool outputs, which represents a surface for indirect prompt injection. 1. Ingestion points: Tool result messages with role='tool' described in SKILL.md. 2. Boundary markers: The basic examples do not include explicit boundary markers or instructions to ignore embedded commands in tool outputs. 3. Capability inventory: The 'liter-llm' client performs network operations to reach LLM providers. 4. Sanitization: The instructional examples do not include explicit sanitization of tool output data, which is standard for basic usage documentation.
Audit Metadata