One-off jobs are nice; workflows that run without babysitting are better. Your vibe-code agent can write scripts, schedule them with cron, and connect Leantime’s API to the other tools your team already lives in. Each workflow below is a small script plus a schedule entry, and the agent builds, tests, and documents both while you watch. Here are ten automations worth setting up once and forgetting.

1. Nightly Overdue Digest

Once you pass fifty open to-dos, nobody holds the full picture in their head. A cron script pulls every overdue to-do through Leantime’s API, groups them by assignee, and emails each person their own list. The agent writes the script, schedules it, and adds logging so failures are visible instead of silent — and it tests one run manually before trusting the schedule. “Write a script that runs every night at 6:30: find all overdue to-dos, group them by assignee, and email each person their own list. Log what it did.”

2. Escalate Stale High-Priority Work

Not all overdue items are equal — a critical bug sitting untouched is a different problem from a low-priority chore. A second check targets high and critical to-dos that are past due and have had no comment in 48 hours: it posts a comment asking for an ETA and emails the project manager a direct list. Since it reads modified dates, the agent can work straight from the database. “Every morning, find high or critical to-dos overdue by more than two days with no comment in the last 48 hours, comment asking for an ETA, and email the PM the list.”

3. Weekly Idea-to-Backlog Sweep

An idea board that never ships anything trains people to stop submitting. Every Monday, the agent ranks open ideas by upvotes, promotes anything above a threshold you set into backlog to-dos linked back to the idea, and posts a comment so voters see what happened. “Every Monday morning, promote ideas with 10 or more votes into the Backlog as to-dos, add a comment on each idea linking the new ticket, and email me what got promoted.”

4. Sync Git Activity to To-Dos

Developers hate updating ticket status, and commit messages already say what’s happening. The agent sets up a webhook listener for your Git host: a commit referencing a to-do ID moves it to In Progress, and one saying “closes” moves it to Done with the commit linked. Have it test the whole chain with a dummy commit before calling it done, since webhook secrets and payload shapes vary between GitHub, GitLab, and Gitea. The status-update ritual disappears. “Set up a webhook listener that moves a to-do to Doing when a branch or commit mentions its ID, and to Done when the commit message says ‘closes’.”

5. Automated Client Status Reports

Clients ask “how’s it going” more often than any owner wants to answer. A Friday cron builds the answer from data: each milestone with its date and percent of to-dos done, the tickets closed this week, and anything at risk — rendered into a clean email per client contact. “Every Friday, generate a status email per active client project: milestones with dates and percent complete, tickets closed this week, and anything at risk. Send it to the client contact on each project.”

6. Timesheet Nagging That Actually Works

Time tracking fails when people forget, not when they refuse. Two small crons solve it: a Friday 3pm reminder to anyone under their weekly hours target, and a Monday morning report of last week’s logged hours per person against target. The agent wires both to Leantime’s timesheets and keeps the tone friendly. “Every Friday at 3pm, remind anyone under 30 logged hours to fill in their timesheet, and Monday morning email me last week’s totals per person.”

7. Support Inbox Into Bug To-Dos

Customer email and your project board should not be separate worlds. A script polls a support mailbox every ten minutes, files each new message as a bug to-do with the sender and full body, tags it by keyword (“error”, “slow”, “can’t log in”), and auto-replies with the ticket number. It deduplicates by subject so reply chains don’t spawn twins. “Poll support@ourcompany.com every 10 minutes, create a bug to-do per new email with the body as the description, and auto-reply to the sender with the ticket number.”

8. Milestone Deadline Watchdog

Milestones slip silently until the day they hit. A daily job checks every milestone’s date and emails the project owner at 14, 7, and 1 days out, each time with the count of open to-dos still attached and the three oldest items. Surprises become a choice instead of an event. “Every morning, check upcoming milestones and email the project owner at 14, 7, and 1 days before the deadline with the open to-do count and the three oldest open items.”

9. Monthly Billable-Hours Export

If you bill clients by the hour, month-end is a data chore. A cron pulls the month’s timesheets per client project, writes one CSV per client, totals the hours, and emails the bundle to you. Tie it to timesheet approval so unapproved time never leaks into an invoice — your bookkeeper will ask which hours are final, and this makes the answer “all of them in the CSV.” “On the first of each month, export last month’s approved timesheets grouped by client into one CSV per client, and email them to me with the totals.”

10. Push Events to Slack or Teams

Chat is where the team actually looks. The agent sets up a script that watches recently changed to-dos — or receives Leantime webhooks where your version supports them — and posts to a channel when something is marked Done or picks up a “blocked” tag, with the project name and a link back. “Post to our #projects channel whenever a to-do moves to Done or gets tagged blocked, with the project name and a link.”

Start with the one automation that solves your most recurring annoyance — for most teams that’s the overdue digest. Each of these is a script plus a cron entry, and your agent builds both while you watch. Leantime on OpenSysLab