The Agent Skills Directory

[COMMAND_EXECUTION]: The skill facilitates the deletion of files on the local file system. It mitigates the risk of unauthorized deletion by requiring the agent to display all targeted paths and obtain an explicit 'yes' from the user before proceeding.
[DATA_EXFILTRATION]: The skill accesses configuration and log files located at ~/.claude/plugins/config/claude-for-legal/legal-builder-hub/install-log.yaml and CLAUDE.md. These files are read to verify that the target skill was installed through the community hub and is eligible for removal.
[PROMPT_INJECTION]: The skill accepts a user-provided 'skill name' argument which determines the file paths to be deleted, creating a surface for indirect prompt injection.
Ingestion points: The [skill name] argument in the uninstall command.
Boundary markers: The skill instructions require the agent to confirm the action with the user and verify the skill against a trusted installation log.
Capability inventory: File system deletion (via the referenced skill-manager) and file write access for appending to the audit log.
Sanitization: The instructions include logic to refuse uninstallation if the target path resolves to a protected first-party plugin directory.

uninstall