ai-local-model-ops
Pass
Audited by Gen Agent Trust Hub on Aug 12, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill consists entirely of instructional markdown, reference tables, and configuration templates. No executable code, malicious scripts, or obfuscation techniques were detected.
- [EXTERNAL_DOWNLOADS]: The skill provides instructions for downloading well-known tools from their official sources, including Ollama, LM Studio, and Microsoft Foundry Local. These are legitimate resources consistent with the skill's stated purpose.
- [REMOTE_CODE_EXECUTION]: A setup recipe includes a template for installing Ollama using
curl | bashfromollama.com. While this pattern is generally noted for caution, it is a standard installation method for this well-known tool and is provided here as documentation for the user to execute manually. - [DATA_EXPOSURE_EXFILTRATION]: The skill addresses the risk of exposing unauthenticated local API endpoints. It provides instructions on how to bind services to a local network (
0.0.0.0) but explicitly labels this as a 'Known Trap' and provides remediation guidance, such as using reverse proxies with TLS and basic authentication. - [INDIRECT_PROMPT_INJECTION]: The skill defines an attack surface by facilitating the connection of local models to external web interfaces (Open WebUI). However, it does not include vulnerable configurations and encourages the use of boundary markers and evaluation sets to mitigate such risks.
Audit Metadata