axolotl

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFE
Full Analysis
  • SAFE (SAFE): The skill is a documentation-centric extension for the Axolotl library. Analysis of the YAML frontmatter, markdown instructions, and index files revealed no malicious intent or security vulnerabilities.
  • Indirect Prompt Injection (LOW): The skill uses a documentation-as-context approach, which is a standard pattern for AI skills. While processing external data (the reference files) is a theoretical injection surface, there are no triggers or executable scripts that would escalate this risk.
  • Ingestion points: references/api.md, references/dataset-formats.md, references/other.md.
  • Boundary markers: None present in the index file; the agent is expected to 'view' these files as needed.
  • Capability inventory: The skill documentation references CLI command execution (e.g., NCCL tests) and Python functions for running commands via cloud providers (run_cmd), but these are examples for the user, not automated skill actions.
  • Sanitization: Not applicable for static documentation content.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 04:57 PM