Provides a simple runner for 'jinx' tasks: invoke a jinx by its registered name or by filesystem path and pass input values as JSON. The skill executes the target jinx and returns either the declared output field or the entire execution context. Useful for iterative development and running ad-hoc agent tasks without leaving the session.
Use when you want to run a small task/script (jinx) during development, test new jinxes before registering them, or execute reusable automation steps by name. Helpful for ML/NLP research workflows where experiments are scripted as jinxes.
Developer and research-focused Python agents that can run repository scripts (e.g., Codex, Copilot, Python-enabled assistants).
This skill has not been reviewed by our automated audit pipeline yet.
