What problem does it solve? Hermes Agent's cron CLI has undocumented quirks — positional prompt ordering, missing --model flags, and script path rules — that cause silent failures, and diagnosing broken jobs requires knowing where output files and delivery origins actually live. ## Core Features & Use Cases - Correct Job Creation: Enforces the critical argument ordering (prompt immediately after schedule) and explains --script, --skill, --deliver, and agent vs no-agent modes. - Failure Diagnosis: Walks through reading timestamped .md output files, grouping errors by root cause (auth expiry, lock contention, timeouts, daemon-in-cron), and fixing causes instead of individual jobs. - Delivery Auditing & Migration: Shows how to read jobs.json origin fields to audit which chats receive job output and how to move jobs between Hermes profiles. - Use Case: When three cron jobs all fail with the same Google OAuth error, use this Skill to trace the .md output logs, identify the shared root cause, and fix the auth once instead of patching three scripts. ## Quick Start Ask the agent to create a weekday 9am cron job that runs a script and delivers results back to the originating chat.