The Agent Skills Directory

Command Execution: The skill performs system operations using git and diff to compare local file states with framework sources. These actions are limited to status reporting and do not involve unauthorized modifications to the system.
External Resource Verification: The skill performs git fetch on repositories defined in the user's configuration to check for updates. The instructions prioritize verifying that the source origin belongs to the project's official infrastructure.
Data Handling and Input Sanitization: The skill processes external information, such as git history and branch names, which are treated as untrusted data. It includes explicit safeguards to prevent the agent from interpreting this data as instructions, mitigating potential indirect prompt injection considerations.
Security Policy Review: The skill audits the configuration file settings.json to ensure that protective measures, such as blocked file paths and network restrictions, are properly configured. This is a read-only audit designed to identify potential regressions in the environment's security posture.

isolated-setup-update