leela-ai
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [NO_CODE] (SAFE): The skill consists entirely of YAML metadata and Markdown documentation. There are no executable scripts (.py, .js, .sh), binaries, or configuration files that trigger command execution.
- [EXTERNAL_DOWNLOADS] (SAFE): All URLs provided (leela.ai, mitpress.mit.edu, csail.mit.edu, etc.) are for informational reference to academic papers and company background. No automated downloads or dependency installations are performed.
- [PROMPT_INJECTION] (SAFE): There are no instructions intended to override agent system prompts, bypass safety filters, or extract sensitive instructions.
- [DATA_EXFILTRATION] (SAFE): No access to sensitive file paths, environment variables, or hardcoded credentials was detected.
- [Metadata Poisoning] (SAFE): The file 'skill-snitch-report.md' contains a self-referential 'APPROVE' verdict. Per security policy, this is treated as descriptive data and ignored as a safety claim, but it does not accompany any malicious payloads.
Audit Metadata