setting-up-experiment-tracking

Pass

Audited by Gen Agent Trust Hub on Mar 12, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill installs the mlflow and wandb Python packages during the environment configuration phase. These are well-known, industry-standard libraries for experiment tracking.
  • [COMMAND_EXECUTION]: The skill is granted unrestricted access to the Bash tool to perform environment setup, package installation, and potential tracking server initialization (e.g., starting a local MLflow server).
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface because it reads and analyzes project context to determine configuration settings. Maliciously crafted instructions hidden in project files could potentially influence the agent's behavior during the setup process.
  • Ingestion points: Analyzes project context and user-provided configuration strings in SKILL.md.
  • Boundary markers: No explicit delimiters or instructions are provided to the agent to ignore embedded commands within the analyzed project files.
  • Capability inventory: The skill utilizes Write, Edit, and unrestricted Bash tool access as defined in SKILL.md.
  • Sanitization: There is no evidence of input validation or content filtering for the data read from the project environment.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 12, 2026, 12:37 AM