Mythos

Objective

**/scheduled** is the management interface for 📝Claude Code's scheduled remote agent triggers — the web dashboard where every trigger you create via the /schedule skill is listed, editable, and runnable on demand. Each trigger spawns a fully isolated remote session in 📝Anthropic's cloud infrastructure on a cron schedule, with its own git checkout, tool allowlist, and optional MCP connectors. The /scheduled dashboard at claude.ai/code/scheduled is the only place you can view all your triggers together, enable or disable them, inspect run history, and delete them — deletion cannot be done from the /schedule skill itself.

Key Facts

  • URL: https://claude.ai/code/scheduled
  • Per-trigger URL: https://claude.ai/code/scheduled/{trigger_id}
  • Execution environment: Anthropic's cloud infrastructure (sandboxed, not user's local machine)
  • Minimum cron interval: 1 hour
  • Creation skill: /schedule in Claude Code
  • Deletion: only via the /scheduled web dashboard (not the skill)
  • MCP access: scoped at trigger creation — must be connected via claude.ai/settings/connectors first
  • Git access: triggers check out specified repos via HTTPS — no authentication to private repos unless a PAT is configured
  • Plan availability: Pro and Max plans

How It Works

Each trigger runs remotely at its cron time. The dashboard shows next run, last run, enabled status, attached git repos, MCP connections, and the full prompt that will be executed. Operators can:

  • Toggle enabled/disabled without losing configuration — useful for pausing a trigger during travel or maintenance windows
  • Edit prompt or cron expression via the dashboard or the /schedule skill
  • Run now to execute a trigger manually for testing, independent of its cron schedule
  • Inspect run history to see past executions and their outputs
  • Delete a trigger permanently (only possible from the dashboard)

The typical lifecycle: create via /schedule skill → manage ongoing via /scheduled dashboard → tune prompts and cadence as behavior proves out.

Why It Matters

/scheduled is the control plane for async Claude Code work. Once you have more than one or two triggers, the dashboard becomes essential for seeing what's running, when, and whether anything has silently failed. Because the skill can't delete or surface run history, the dashboard is the operator's source of truth. It also exposes the trigger ID required for /schedule run and update operations.

FAQ

What is the Claude Code /scheduled page?

/scheduled is the web dashboard at claude.ai/code/scheduled where all scheduled remote agent triggers are listed and managed. It complements the /schedule skill in Claude Code — the skill creates triggers, the dashboard manages them.

How do I delete a scheduled trigger?

Deletion is only possible from the /scheduled dashboard. Open the trigger detail page and use the delete action. The /schedule skill does not support deletion.

Can scheduled triggers access local files?

No. Triggers run in Anthropic's cloud with their own sandboxed environment. They can read files from attached git repositories and use whatever tools are allowlisted at creation time, but they have no access to the user's local filesystem, local services, or local environment variables.

How do I attach MCP servers to a trigger?

Connect the MCP server at claude.ai/settings/connectors first. Then when creating the trigger via /schedule, reference the connector by connector_uuid and name. Triggers created without MCP connections cannot be retroactively granted them without an update operation.

Where does a trigger's output go?

Wherever the prompt directs it. Common patterns: open a GitHub issue, create a MythOS memo (with MCP), send a notification, push a branch. The trigger's prompt must explicitly instruct the agent to produce reviewable output — there is no automatic surface for session transcripts.

Related

Subjective

Contexts

Created with 💜 by One Inc | Copyright 2026