youmind-deep-research

Pass

Audited by Gen Agent Trust Hub on Mar 17, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill installs the official vendor CLI package @youmind-ai/cli globally using npm. This is a standard requirement for the skill to function and originates from the authorized vendor.
  • [COMMAND_EXECUTION]: The skill executes several bash commands using the youmind CLI to interact with the research API. It also uses node -e to parse and process JSON results from the command line, which is a safe practice for data extraction in this context.
  • [COMMAND_EXECUTION]: User-provided research topics are interpolated into JSON strings within bash commands. While this presents a minor risk of command injection if the agent does not correctly escape special characters (like single quotes), it is a common pattern in agent skills and typically handled by the platform's tool execution layer.
  • [PROMPT_INJECTION]: The skill processes external data (research results) which could potentially contain indirect prompt injections.
  • Ingestion points: Data is fetched via youmind call listMessages in SKILL.md (Step 5).
  • Boundary markers: None explicitly defined in the prompt templates for result summarization.
  • Capability inventory: The skill has access to bash and node execution.
  • Sanitization: Results are parsed as JSON via node before being presented to the LLM for summarization.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 17, 2026, 01:58 PM