apple-reminders
Pass
Audited by Gen Agent Trust Hub on Sep 1, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes the
remindctlbinary to interact with the local Reminders application on macOS, executing commands to create, view, and modify task data. - [INDIRECT_PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection by passing untrusted user data into shell command arguments. * Ingestion points: User-provided strings for task titles, list names, and dates are ingested through the
addandlistcommands inSKILL.md. * Boundary markers: Although the skill documentation uses double quotes for task titles (e.g.,remindctl add "Buy milk"), it does not provide explicit instructions for the agent to escape shell metacharacters or subshell invocations. * Capability inventory: The skill has the capability to read all reminders, create new tasks, and delete existing entries, which requires user-granted access to the Reminders database. * Sanitization: The skill instructions do not specify any sanitization, filtering, or validation steps for the user-supplied content before it is executed. - [EXTERNAL_DOWNLOADS]: The skill requires the external installation of the
remindctltool from a third-party Homebrew tap (steipete/tap/remindctl). This is a standard and documented method for acquiring the tool from its developer's repository.
Audit Metadata