daily-update has two modes. Run mode checks each ingested source against the manifest and marks it fresh, stale or missing, refreshes index.md and hot.md through the obsidian-wiki command, runs wiki-lint on a schedule, writes a state file and reports which sources need syncing.
Setup mode installs the schedule for 9 AM: a launchd agent on macOS, a systemd user timer on Linux, or a cron line otherwise, plus an optional terminal reminder when the wiki is stale.
What You Get
- A freshness report showing fresh, stale and missing sources with the ingest command to run for each
- A reconciled index and a refreshed hot cache
- A weekly health check from wiki-lint, configurable with LINT_SCHEDULE
- A scheduled daily run set up for launchd, systemd or cron, with an optional terminal notification
Common Use Cases
- Run a quick morning refresh of the vault
- Set up a scheduled daily maintenance job
- Find out which agent histories have new sessions to ingest
- Get a terminal reminder when the wiki has gone stale
What Makes It Unique
It picks the right scheduler for the platform and scopes all runtime state to the vault, so several vaults can run side by side. It also hands its own result to a validator subagent, which checks timestamps, counts and the hot cache date before the cycle is logged.
Under The Hood
The skill is instructions for the agent. A run reads the vault manifest, calls the obsidian-wiki memory commands, writes small state files in a per-vault config folder, appends a log line and may launch wiki-lint and the impl-validator subagent. Setup mode goes further: it runs shell commands to write a launchd plist, systemd unit files or a crontab entry, can append a source line to your shell rc file after asking, and runs the daily-update script once. It also refreshes a QMD search index when one is configured. It makes no web requests.