The Agent Skills Directory

[COMMAND_EXECUTION]: The skill facilitates performance optimization by executing shell commands via the omx CLI tool and user-defined evaluator commands such as npm run perf:startup. These executions are central to the skill's utility in benchmarking and verifying system performance.
[INDIRECT_PROMPT_INJECTION]: The skill defines a workflow for processing untrusted data, specifically the performance objective and evaluator contract. It mitigates potential risks by requiring a 'Completion Gate' where a passing evaluator checkpoint and a completion audit must be documented before the agent is permitted to mark a goal as complete.

performance-goal