Schedule a CONTINUATION of your own work to run later — the way to handle recurring or deferred tasks instead of looping in-process or busy-waiting. Pick exactly one cadence: when (one-shot RFC3339 UTC), interval_minutes (every N minutes), or cron (5-field expr; supports */N steps and @daily/@hou...
AI agents use schedule_self to create or update resources in Yaver — usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your Yaver environment.
| Parameter | Type | Required | Description |
|---|---|---|---|
cron | string | — | 5-field cron expression (minute hour day month weekday), e.g. '*/30 * * * *' or '@daily'. Use this OR when OR interval_minutes. |
memo | string | — | Optional notes carried verbatim into the next run's prompt (state, findings, where you left off). |
when | string | — | One-shot run time, RFC3339 UTC (e.g. 2026-06-17T09:00:00Z). Use this OR interval_minutes OR cron. |
model | string | — | Optional model override for the next run. |
title | string | — | Optional label. Defaults to a truncation of prompt. |
prompt | string | Yes | The instruction the next run executes. Self-contained — the next process has no memory of the current turn. |
resume | boolean | — | Recurring only: natively resume the previous run's session each fire (claude/glm/codex by session id, opencode by --continue) instead of starting cold. Default |
runner | string | — | Runner for the next run: claude | codex | opencode | glm. Defaults to this agent's default runner. |
max_runs | integer | — | Stop after this many fires (0 = use the 100-fire safety cap for recurring; one-shot ignores this). |
interval_minutes | integer | — | Repeat every N minutes (minimum 1). Use this OR when OR cron. |
Parameters from the server's own tool schema.
An AI agent can call schedule_self faster than any human can review — one bad instruction and it creates or modifies resources in Yaver by the hundred, each call as confident as the last.
Risk signalsHigh parameter count (10 properties) · Bulk/mass operation — affects multiple targets
Attacks that exploit this kind of access
Schedule a CONTINUATION of your own work to run later — the way to handle recurring or deferred tasks instead of looping in-process or busy-waiting. Pick exactly one cadence: when (one-shot RFC3339 UTC), interval_minutes (every N minutes), or cron (5-field expr; supports */N steps and @daily/@hourly macros). The next run starts as a FRESH process (no memory of this turn) so put everything it needs into prompt and memo. memo is carried verbatim into the next run's prompt. runner defaults to this agent's default; pass it to pin claude/codex/opencode/glm. Recurring schedules without max_runs are capped at 100 fires. It is categorised as a Write tool in the Yaver MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
schedule_self accepts 10 parameters: cron, memo, when, model, title, prompt, resume, runner, max_runs, interval_minutes. Required: prompt. The full parameter table on this page comes from the server's own tool schema.
Register the Yaver MCP server in PolicyLayer and add a rule for schedule_self: allow, deny, rate-limit, or require approval. Point your MCP client at the PolicyLayer proxy URL and the rule is enforced on every call, before it reaches Yaver. Nothing to install.
schedule_self is a Write tool with medium risk. Write tools should be rate-limited to prevent accidental bulk modifications.
Yes. Add a rate_limit block to the schedule_self rule in your PolicyLayer policy. For example, setting max: 10 and window: 60 limits the tool to 10 calls per minute. Rate limits are tracked per agent session and reset automatically.
Set action: deny in the PolicyLayer policy for schedule_self. The AI agent will receive a policy violation error and cannot call the tool. You can also include a reason field to explain why the tool is blocked.
schedule_self is provided by the Yaver MCP server (yaver-cli). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.