schedule

Manage cron-based scheduled tasks that deliver messages to the tinyclaw queue via the API server. Each schedule fires at a cron interval and POSTs a routed message (@agent_id <task> ) to POST /api/message , where the queue processor picks it up and invokes the target agent.

Safety Notice

This listing is imported from skills.sh public index metadata. Review upstream SKILL.md and repository scripts before running.

Copy this and send it to your AI assistant to learn

Install skill "schedule" with this command: npx skills add tinyagi/tinyclaw/tinyagi-tinyclaw-schedule

Schedule Skill

Manage cron-based scheduled tasks that deliver messages to the tinyclaw queue via the API server. Each schedule fires at a cron interval and POSTs a routed message (@agent_id <task> ) to POST /api/message , where the queue processor picks it up and invokes the target agent.

Commands

Use the bundled CLI scripts/schedule.sh for all operations.

Create a schedule

scripts/schedule.sh create
--cron "EXPR"
--agent AGENT_ID
--message "Task context for the agent"
[--channel CHANNEL]
[--sender SENDER]
[--label LABEL]

  • --cron — 5-field cron expression (required). Examples: "0 9 * * " (daily 9am), "/30 * * * *" (every 30 min), "0 0 * * 1" (weekly Monday midnight).

  • --agent — Target agent ID (required). Must match an agent configured in settings.json.

  • --message — The task context / prompt sent to the agent (required).

  • --channel — Channel name in the queue message (default: schedule ).

  • --sender — Sender name in the queue message (default: Scheduler ).

  • --label — Unique label to identify this schedule (default: auto-generated). Use a descriptive label for easy management.

List schedules

scripts/schedule.sh list [--agent AGENT_ID]

Lists all tinyclaw schedules. Optionally filter by --agent to show only schedules targeting a specific agent.

Delete a schedule

scripts/schedule.sh delete --label LABEL scripts/schedule.sh delete --all

Delete a specific schedule by label, or delete all tinyclaw schedules.

Workflow

  • Confirm the target agent ID exists (check settings.json or ask the user).

  • Determine the cron expression from the user's description (e.g., "every morning" → "0 9 * * *" ).

  • Compose a clear task message — this is the prompt the agent will receive.

  • Run scripts/schedule.sh create with the parameters.

  • Verify with scripts/schedule.sh list .

Cron expression quick reference

┌───────────── minute (0-59) │ ┌───────────── hour (0-23) │ │ ┌───────────── day of month (1-31) │ │ │ ┌───────────── month (1-12) │ │ │ │ ┌───────────── day of week (0-7, 0 and 7 = Sunday) │ │ │ │ │


Pattern Meaning

0 9 * * *

Daily at 9:00 AM

0 9 * * 1-5

Weekdays at 9:00 AM

*/15 * * * *

Every 15 minutes

0 */2 * * *

Every 2 hours

0 0 * * 0

Weekly on Sunday midnight

0 0 1 * *

Monthly on the 1st

30 8 * * 1

Monday at 8:30 AM

Examples

Daily report

scripts/schedule.sh create
--cron "0 9 * * *"
--agent analyst
--message "Generate the daily metrics report and post a summary"
--label daily-report

Periodic health check

scripts/schedule.sh create
--cron "*/30 * * * *"
--agent devops
--message "Run health checks on all services and report any issues"
--label health-check

Weekly code review reminder

scripts/schedule.sh create
--cron "0 10 * * 1"
--agent coder
--message "Review open PRs and summarize status"
--label weekly-pr-review

List and clean up

See all schedules

scripts/schedule.sh list

See only schedules for @coder

scripts/schedule.sh list --agent coder

Remove one

scripts/schedule.sh delete --label health-check

Remove all

scripts/schedule.sh delete --all

How it works

  • Schedules are stored as system cron entries tagged with # tinyclaw-schedule:<label> .

  • When a cron job fires, it POSTs a message to the API server (POST /api/message ) with the @agent_id routing prefix.

  • The queue processor picks up the message and invokes the target agent, exactly like a message from any channel.

  • Responses are stored in the SQLite queue and delivered by channel clients.

For queue message format details, see references/queue-format.md .

Source Transparency

This detail page is rendered from real SKILL.md content. Trust labels are metadata-based hints, not a safety guarantee.

Related Skills

Related by shared tags or category signals.

Automation

agent-browser

No summary provided by upstream source.

Repository SourceNeeds Review
General

send-user-message

No summary provided by upstream source.

Repository SourceNeeds Review
General

skill-creator

No summary provided by upstream source.

Repository SourceNeeds Review
General

imagegen

No summary provided by upstream source.

Repository SourceNeeds Review