The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill's function is to analyze externally provided code implementations, which serves as a primary ingestion point for untrusted data. * Ingestion points: The skill analyzes code for algorithmic complexity, database queries, and memory management based on user implementation. * Boundary markers: The instructions lack explicit boundary markers or warnings to the model to ignore natural language instructions that may be hidden in code comments or strings. * Capability inventory: The skill is focused on analysis but operates within an agent context that includes a Task tool for launching further actions, as seen in the examples. * Sanitization: There are no defined mechanisms for the agent to sanitize or validate the content of the code before processing it.

performance-oracle