lecun-world-model
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill consists entirely of natural language instructions for a reasoning framework. No executable scripts, command-line operations, or file system modifications are present in the skill package.
- [PROMPT_INJECTION]: The instructions direct the agent to adopt a specific persona ('AI researcher') and follow a structured nine-step reasoning process. These are standard instructional techniques designed to influence reasoning depth and do not include attempts to bypass safety filters or disregard core system instructions.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to analyze user-provided scenarios. Ingestion point: User-supplied situations via the SKILL.md interaction logic. Boundary markers: None present. Capability inventory: No tools, network operations, or file-system capabilities are requested or used across the skill. Sanitization: None present. Because the skill has no actionable capabilities to exploit, this attack surface presents no practical security risk.
- [METADATA_POISONING]: The metadata fields (name and description) accurately reflect the skill's purpose and do not contain hidden instructions or deceptive content.
Audit Metadata