apple-reminders-hardened

Pass

Audited by Gen Agent Trust Hub on Apr 21, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADS
Full Analysis
  • [EXTERNAL_DOWNLOADS]: Fetches the remindctl binary via Homebrew from a third-party repository (steipete/tap/remindctl) or allows for manual compilation from source via pnpm.- [DATA_EXFILTRATION]: Contains explicit defensive instructions and guardrails to prevent data exfiltration by forbidding the piping of reminder data to network-transmitting commands.- [PROMPT_INJECTION]: Includes specific instructions designed to resist adversarial framing and ensure user confirmation for destructive actions.- [PROMPT_INJECTION]: The skill processes untrusted reminder content which could facilitate indirect prompt injection. 1. Ingestion points: remindctl command output from local database. 2. Boundary markers: Absent. 3. Capability inventory: remindctl delete, edit, and list operations. 4. Sanitization: Absent.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 21, 2026, 06:30 PM
Security Audit — agent-trust-hub — apple-reminders-hardened