learning-language-level-calibration
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- [Data Exposure & Exfiltration] (SAFE): No network operations or sensitive file access patterns detected.
- [Remote Code Execution] (SAFE): No remote code downloads or dynamic execution of untrusted input.
- [Indirect Prompt Injection] (SAFE): Analysis of attack surface: 1. Ingestion points: 'source_content' parameter in skill.py. 2. Boundary markers: None present. 3. Capability inventory: Minimal; the skill returns a static dictionary and does not perform file writes, network requests, or command execution. 4. Sanitization: None present. Due to the lack of capabilities, this surface is not exploitable.
- [Metadata Poisoning] (SAFE): The functionality described in SKILL.md (CLI flags) is not implemented in the python script, representing a functional stub rather than a security threat.
Audit Metadata