llm-config
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill provides an interface for adapting LLM behavior using external data sources, creating a potential indirect prompt injection surface.
- Ingestion points: The skill ingests training data for MicroLoRA adaptation (Step 4) and feedback signals for SONA adaptation (Step 6) as described in SKILL.md.
- Boundary markers: The instructions lack specific delimiters or instructions to ignore embedded malicious content within the adaptation data.
- Capability inventory: The skill has access to the Bash tool and several specialized model adaptation plugins (ruvllm_microlora_adapt, ruvllm_sona_adapt).
- Sanitization: No data validation or sanitization mechanisms are defined for the external inputs used in the adaptation process.
Audit Metadata