bdistill-behavioral-xray
Warn
Audited by Gen Agent Trust Hub on Oct 28, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The instructions require the user to install an external Python package named
bdistill. This package is not associated with any recognized trusted vendors or well-known services. - [COMMAND_EXECUTION]: The skill provides shell commands for the agent or user to execute, specifically for package installation (
pip install bdistill) and for adding an MCP server (claude mcp add bdistill). - [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze output generated by AI models, which serves as a surface for indirect prompt injection.
- Ingestion points: Responses generated by the model during the 30-question behavioral probe described in
SKILL.md. - Boundary markers: There are no specified delimiters or instructions to treat the model's responses as untrusted data.
- Capability inventory: The skill uses shell command execution and MCP tool integration.
- Sanitization: The instructions do not include any steps for sanitizing, escaping, or validating the ingested model responses before analysis.
Audit Metadata