Skills
skills/openclaw/skills/litellm/Gen Agent Trust Hub

litellm

Pass

Audited by Gen Agent Trust Hub on Mar 2, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The skill provides a pathway for LLM calls, creating a surface for indirect prompt injection when handling external data.
  • Ingestion points: The prompt and system arguments in scripts/llm_call.py accept arbitrary text input.
  • Boundary markers: The message list is constructed in scripts/llm_call.py without using delimiters or instructions to the model to ignore potential commands within the user input.
  • Capability inventory: The script executes network calls to various model provider endpoints using the litellm library.
  • Sanitization: No validation or sanitization is applied to the input strings before they are transmitted to the LLM API.
  • [EXTERNAL_DOWNLOADS]: The skill documentation refers to the litellm library, which is a well-known package for LLM orchestration.
  • Evidence: SKILL.md contains instructions for installing the library via pip install litellm.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 2, 2026, 10:42 AM