Slash Commands

Slash commands are shortcuts you type in the Agenties chat input to navigate, configure, and control your agent workspace. Type a / in the chat bar to see an autocomplete menu, or use/help for a full list.

Tip:Custom slash commands can be added by placing skill files in ~/.claude/commands/. The heartbeat self-improvement agent can write new commands here automatically. See the Heartbeats docs for details.

Jump to different sections of the Agenties UI.

Command	Arguments	Description
`/activity`	—	Open the Activity feed — a timeline of all agent events, spawns, completions, and sync events.
`/agents`	—	Open the Agents panel — live view of all running and recent agent processes.
`/chat`	—	Return to the main chat with the orchestrator.
`/dashboard`	—	Open the project dashboard — summary of open issues, active agents, and recent activity.
`/heartbeats`	—	Open the Heartbeats panel — configure and view scheduled agent spawns.
`/history`	—	Open the message history — full mailbox timeline with type filters.
`/issues`	—	Open the issue tracker.
`/manager`	—	Open the Intelligent Manager config — tune the orchestrator's system prompt and behaviour.
`/routines`	—	Open the Routines panel — manage schedule, API, and GitHub automation routines.
`/settings`	—	Open Settings.
`/team`	—	Open the Team tab — manage your agent roster.

Configuration

Adjust the current session's model, effort, and other settings.

Command	Arguments	Description
`/model`	<model-id>	Set the model for the current orchestrator session. Example: /model claude-opus-4-7. Takes effect on the next orchestrator turn.
`/effort`	<level>	Set the effort level for the current session: off, low, medium, high, xhigh, or max.
`/memory`	—	Show the current shared state, vault connection status, and last 3 session summaries.
`/permissions`	—	Open the permissions manager — view auto-allowed tools and permission history.
`/mcp`	—	Show the status of all connected MCP servers and their available tools.

Information

Query usage, costs, and project status.

Command	Arguments	Description
`/cost`	—	Show token usage and estimated cost for the current session and the last 7 days, broken down by model.
`/status`	—	Show a compact status summary: active agents, open issues, current phase, and sync status.
`/doctor`	—	Run a health check: verify Claude Code CLI, MCP server, vault connection, and sync status. Suggests fixes for any issues found.
`/usage`	—	Detailed usage statistics: tokens per agent role, cost per team member, tasks per day.
`/tokens`	—	Show the current context window usage for the orchestrator session — useful for deciding when to /wrapup.

Session

Manage the orchestrator session, export data, and control context.

Command	Arguments	Description
`/clear`	—	Clear the current orchestrator session and start fresh. Shared state and vault are preserved. The old session ID is discarded.
`/export`	[markdown\|json]	Export the current chat history. Default: markdown. JSON export includes full message metadata.
`/retry`	—	Resend the last message. Useful if the orchestrator produced a poor response or timed out.
`/init`	—	Initialise context: load CONTEXT.md, decisions.md, and recent session notes into the orchestrator. Run at the start of each new session.
`/wrapup`	—	Close the session: write a structured summary to vault/sessions/YYYY-MM-DD.md, upload to NotebookLM, and start a fresh orchestrator session.
`/projects`	[N]	List recent projects. N defaults to 5. Click a project name to switch to it.
`/vault`	[path]	Browse the project vault. If a path is provided, show that note's content directly in the chat.

General

Utility commands.

Command	Arguments	Description
`/help`	[command]	Show all available commands, or detailed help for a specific command.
`/stop`	—	Stop all running agents immediately. Triggers the BFS cancel cascade. Equivalent to clicking the Stop button in the UI.

Custom commands

You can create your own slash commands by adding Markdown files to ~/.claude/commands/. The filename (without .md) becomes the command name. The file content is the system prompt for that command.

~/.claude/commands/sprint-review.md

---
description: Generate a sprint review summary from this week's done issues
---

Read all issues with status=done that were closed this week using list_issues.
For each issue, summarise what was implemented.
Write a sprint review document to vault/reports/YYYY-WW-sprint.md.
Post a brief summary to the chat when done.

This creates a /sprint-review command. Type it in the chat and the orchestrator runs the command as a new turn with the custom prompt injected.