The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it processes data from external sources (Kaggle).
Ingestion points: Data enters the agent's context through tool execution results and RUBE_SEARCH_TOOLS responses.
Boundary markers: No boundary markers (e.g., XML tags or delimiters) are specified to separate untrusted data from instructions.
Capability inventory: The skill can perform significant actions via RUBE_MULTI_EXECUTE_TOOL and RUBE_REMOTE_WORKBENCH (which facilitates running composio tools).
Sanitization: There is no evidence of sanitization or validation performed on data retrieved from Kaggle before it is processed by the agent.
[SAFE]: The skill references vendor-specific resources including rube.app (MCP server endpoint) and composio.dev (toolkit documentation). These are standard components of the author's (composiohq) ecosystem.
[DYNAMIC_EXECUTION]: The skill uses RUBE_SEARCH_TOOLS to discover tool schemas at runtime and RUBE_MULTI_EXECUTE_TOOL to execute them. This dynamic discovery is a standard part of the MCP (Model Context Protocol) workflow for this vendor.

kaggle-automation