confident-eval-setup
Pass
Audited by Gen Agent Trust Hub on Oct 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external, untrusted content from source code and dataset files (CSV, JSON) to automate the configuration of AI connections and datasets.
- Ingestion points: Workflow steps 3 and 6 in
SKILL.md, and the dataset building section inreferences/datasets.md. - Boundary markers: The skill does not define specific delimiters for separating ingested code or data from instructions.
- Capability inventory: The skill uses MCP tools like
create_ai_connectionandpush_datasetto perform network-based operations. - Sanitization: No explicit sanitization of ingested content is performed, though the instructions explicitly direct the agent to ask the user for secrets and avoid printing them.
- [EXTERNAL_DOWNLOADS]: The skill references and connects to the vendor's official MCP server at
mcp.confident-ai.com. - [COMMAND_EXECUTION]: The skill performs project management and configuration tasks using the Confident AI MCP server tools.
Audit Metadata