The Agent Skills Directory

[SAFE]: The skill's core functionality is to maintain a configuration file located at ~/.cursor/rules/pstack-models.mdc. This file defines model overrides for the 'pstack' tool, allowing the user to customize which AI models are used for various development roles.
[COMMAND_EXECUTION]: The skill instructions direct the agent to detect available models using the environment's CLI or API. This is a legitimate discovery step to ensure the configuration remains valid and does not point to unavailable models.
[DATA_EXPOSURE]: While the skill accesses the filesystem to read and write configuration, it is restricted to the specific .cursor/rules directory intended for this purpose. No sensitive data access or network exfiltration patterns were detected.

setup-pstack