mixpanelyst

Warn

Audited by Gen Agent Trust Hub on Oct 12, 2026

Risk Level: MEDIUMDYNAMIC_CONTEXT_INJECTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [DYNAMIC_CONTEXT_INJECTION]: The skill utilizes shell command execution at load time within SKILL.md to verify the environment setup and retrieve help documentation for the Mixpanel CLI tool.- [DYNAMIC_EXECUTION]: The primary workflow of the skill involves the automated generation of Python scripts which are then executed using a specific local interpreter to perform complex data analysis and visualization.- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external Mixpanel projects, including event properties and business documentation. Ingestion points: Untrusted content is ingested through methods like ws.get_business_context(), ws.query(), and ws.list_metrics() as documented in references/business-context.md and references/exploration.md. Boundary markers: The skill does not implement explicit delimiters or instructional warnings to isolate processed data from the agent's control logic. Capability inventory: The agent possesses capabilities for shell command execution (Bash) and file system modifications (Write, Edit) as defined in the allowed-tools metadata. Sanitization: There is no evidence of data sanitization or validation of external content before it is processed or used to influence agent behavior.- [COMMAND_EXECUTION]: The skill is configured to allow the execution of arbitrary commands through uv run * and specific paths for the Python interpreter and Mixpanel CLI, facilitating its core analysis functionality.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Oct 12, 2026, 03:24 PM