The Agent Skills Directory

[SAFE]: The skill's primary function is to guide an AI agent in reviewing cursor trajectory and keyboard focus prefetching logic within project files. It provides specific audit rules (e.g., prefetch-trajectory-over-hover) to identify sub-optimal performance patterns.
[REMOTE_CODE_EXECUTION]: The skill explicitly instructs the agent not to add new dependencies or install libraries during the review process, mitigating supply chain or remote execution risks. It mentions the useForesight library only as a reference for existing project patterns.
[DATA_EXFILTRATION]: There are no network operations, credential access, or sensitive file reads performed by the skill itself. The auditing process is confined to the local files provided in the agent's scope.
[PROMPT_INJECTION]: The instructions provided to the agent are focused on the technical audit task and do not contain attempts to override safety filters, bypass instructions, or extract system prompts.

predictive-prefetch