evaluate-rag
Pass
Audited by Gen Agent Trust Hub on Mar 3, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill is entirely instructional and does not contain any executable scripts or binary files.
- [NO_CODE]: No code-bearing files such as Python scripts (.py), JavaScript (.js), or shell scripts (.sh) are present. It contains only a markdown guide.
- [PROMPT_INJECTION]: The content was analyzed for prompt injection patterns. No instructions were found that attempt to bypass AI safety filters, override system constraints, or extract system prompts.
- [DATA_EXFILTRATION]: No network-capable commands (e.g., curl, wget) or sensitive file path references (e.g., .ssh, .aws) were detected.
- [EXTERNAL_DOWNLOADS]: The skill does not define any external dependencies or remote script downloads.
- [INDIRECT_PROMPT_INJECTION]: While the skill provides a template for processing text chunks (untrusted data), it lacks any associated executable capabilities (tools, scripts, or network access) that could be exploited through such an injection, making the risk negligible.
Audit Metadata