[FEATURE] Expose /clear and /compact as callable tool actions

Resolved 💬 2 comments Opened Mar 17, 2026 by skalaydzhiyski Closed Apr 15, 2026

Preflight Checklist

  • [x] I have searched existing requests and this feature hasn't been requested yet
  • [x] This is a single feature request (not multiple features)

Problem Statement

/clear and /compact are handled by the CLI's UI layer and cannot be triggered by MCP tools, hooks, or the agent itself. This means users must manually type slash commands to manage their conversation — the agent has no way to act on natural language requests like "please compact your conversation" or "clear the history." This creates a gap between the UI layer and the agent layer, limiting the ability to build conversation management workflows through MCP servers or automation.

Proposed Solution

Expose clear and compact as agent-layer primitives that can be invoked via tool calls. This would allow:

  • MCP tools to trigger conversation management actions - The agent to respond to natural language requests to compact or clear
  • Optional parameters (e.g., compact focus instructions) to be passed programmatically

By Having that you can infinitely long sessions running without having to use a files and other clunky approaches. The agent can in a sense "forget" things in a controlled manner which is greatly powerful for research and long running tasks in general.

Alternative Solutions

_No response_

Priority

High - Significant impact on productivity

Feature Category

CLI commands and flags

Use Case Example

_No response_

Additional Context

_No response_

View original on GitHub ↗

This issue has 2 comments on GitHub. Read the full discussion on GitHub ↗