Skip to content

Commands Reference

CoClaw provides commands through two interfaces: slash commands in Copilot Chat and Command Palette commands.

Chat Slash Commands

Use these in Copilot Chat by typing @CoClaw /command.

/memory — Show Memory

Displays a summary of what CoClaw currently remembers:

  • Long-term entries — Top 20 entries sorted by importance, showing type, importance score, and content.
  • Daily log — Top 10 entries from today's session.

Usage:

@CoClaw /memory

/distill — Distill to Long-Term

Triggers distillation of today's daily log entries into curated long-term memory. The model consolidates, deduplicates, and prioritizes entries.

Usage:

@CoClaw /distill

After distillation, CoClaw reports the number of entries created. Automatic deduplication runs afterward.

/clear — Clear Session Memory

Clears today's daily log. Long-term memory is preserved.

Usage:

@CoClaw /clear

/soul — Edit Identity

Opens the SOUL.json file for in-line editing of CoClaw's identity and behavior. See Identity & Profile for details.

Usage:

@CoClaw /soul

/auto — Start Telegram Bridge

⚠️ Deprecated: /auto will be removed in version 1.0.0. Use /open instead.

Starts the Telegram polling bridge. CoClaw begins listening for messages from your linked Telegram bot and processes them with full tool access.

Usage:

@CoClaw /auto

Requires a linked Telegram bot. See Telegram Bridge for setup instructions.

/open — Start OpenClaw Mode

Starts OpenClaw mode, which extends the Telegram bridge with:

  • Workspace-backed MEMORY.md integration
  • Daily logs in the workspace
  • Heartbeat checks
  • Cron job support

Usage:

@CoClaw /open

Use this mode when you want a more persistent, proactive Telegram workflow.

/agents — Multi-Agent Orchestration

Runs your task through the multi-agent orchestrator. A Planner produces a JSON DAG of subtasks, the Coder agent (or several in parallel) implements them, the Reviewer reads the result, the Tester adds tests, and the Memory agent persists facts. Coder tasks are auto-fanned out across multiple parallel agents based on file paths and concern lanes (frontend / backend / data / tests / docs / etc.).

Usage:

@CoClaw /agents Build a login page with API and tests

Live progress is shown in the Agents sidebar (the CoClaw activity bar icon). Tune fan-out width with CoClaw.agents.maxParallelCoders (ceiling) and CoClaw.agents.minParallelCoders (floor — pads small tasks with generic lanes), and switch modes with CoClaw.agents.mode (off, slash, always).

Every /agents run also ends with an automatic final reviewer task (tagged [final review] in the plan) that depends on every coder and tester sibling, reads their shared-memory output, and produces a single consolidated review with an APPROVED / CHANGES_REQUESTED verdict. This is controlled by CoClaw.agents.finalReviewer (default auto — skipped when the planner already produced a covering reviewer; set to always to force one on every run, or off to disable). The final reviewer uses the hard tier, so it picks up whatever model you've assigned to CoClaw.models.hard.

Command Palette Commands

Open the Command Palette with Ctrl+Shift+P and type the command name.

CoClaw: Select Model

Opens a quick-pick dialog showing all available Copilot models. Displays each model's name, family, and maximum input tokens. Your selection is persisted across sessions.

CoClaw: Browse Memory

Opens the Memory Browser — a webview panel where you can:

  • View all long-term and daily log entries.
  • Search and filter entries by type or keyword.
  • Edit entry content and importance.
  • Pin/unpin entries to protect them from decay and pruning.
  • Promote daily log entries to long-term memory.
  • Delete individual entries.

CoClaw: Clear Session Memory

Shows a modal dialog with three options:

  • Daily Only — Clear today's daily log.
  • Long-Term Only — Clear all long-term entries.
  • All Memory — Clear everything.

CoClaw: Edit Identity (SOUL)

Opens SOUL.json in the editor. This file controls CoClaw's name, role, instructions, and tone. Changes take effect on the next chat message.

CoClaw: Edit Profile (USER)

Opens USER.json in the editor. This file stores your coding preferences (language, style, frameworks, etc.). Changes take effect on the next chat message.

CoClaw: Distill to Long-Term Memory

Same as the /distill slash command — consolidates daily logs into long-term memory using the active model.

CoClaw: Import Memories

Opens a file picker for JSON files. Imports an array of memory entries, validating each entry's type, content length, and tags.

CoClaw: Export Memories

Opens a save dialog. Exports all memories (long-term + daily) as a JSON array.

CoClaw: Deduplicate Memory

Scans both memory layers for near-duplicate entries (>60% word overlap) and removes the less important duplicates.

CoClaw: Stop Response

Cancels the currently active CoClaw response. Useful if the assistant is stuck in a long tool-use loop.

Two-step setup process:

  1. Enter your Bot Token (from @BotFather on Telegram).
  2. Enter your User ID (numeric, from @userinfobot).

After linking, use /auto in chat to start the bridge. See Telegram Bridge.

Shows a confirmation dialog, then stops the bot and removes stored credentials.

CoClaw: Clear All Cron Jobs

Removes all persisted cron jobs from VS Code storage. If OpenClaw mode is currently running, active in-memory cron jobs are also cleared immediately.

CoClaw: Open Cron Storage

Opens the persisted cron storage folder in your operating system's file explorer. Useful for debugging or manual inspection.

CoClaw: Open Settings

Opens VS Code settings filtered to the CoClaw.* section for quick access to all configuration options.

Follow-Up Suggestions

After each chat response, CoClaw offers two follow-up actions:

  1. "Check memory" — Runs /memory to see what was remembered.
  2. "Distill to long-term" — Runs /distill to save facts to long-term storage.