clawreverse
Inspect, checkpoint, rollback, and branch OpenClaw sessions with the ClawReverse plugin. Use when a user wants to recover from bad tool or file changes, restore a known-good checkpoint, inspect session lineage, create a child branch from an earlier point, or avoid re-spending tokens after a long OpenClaw run. Do not use for plain git history operations or non-OpenClaw workspaces.
Install via CLI (Recommended)
clawhub install openclaw/skills/skills/binhuangpjlab/clawreverse-skillWhat This Skill Does
ClawReverse is a sophisticated session management plugin for the OpenClaw ecosystem, designed to provide granular control over agent task history. By introducing the openclaw reverse command family, it allows users to inspect, checkpoint, rollback, and branch their OpenClaw sessions. This skill acts as a time-travel mechanism for AI agents, enabling the preservation of complex task states. It effectively captures the agent's internal state before any destructive operations, ensuring that if a tool call produces unexpected results or if the task veers off course, the agent can revert to a 'known-good' state without discarding the progress made leading up to that point.
Installation
To integrate ClawReverse, first verify your current installation by running openclaw reverse --help. If the command is unrecognized, execute openclaw plugins install -l "{baseDir}" to load the plugin bundle. Once installed, initialize the configuration with openclaw reverse setup. For custom environments, use openclaw reverse setup --base-dir /path/to/openclaw-state to ensure the plugin targets the correct directory. Finally, restart the OpenClaw Gateway to apply these changes. This setup process creates a local manifest for all session checkpoints, ensuring metadata is persisted across agent restarts.
Use Cases
ClawReverse is invaluable for long-running agentic workflows. Use it when you need to undo a specific tool call sequence that resulted in unintended file modifications. It is particularly effective for branching: if you want to test two different solution strategies for a complex bug, you can create a checkpoint at the current state and spawn a child branch for each experiment, keeping the parent session pristine. Additionally, it prevents token waste by allowing you to resume complex analysis from a specific checkpoint rather than re-executing the entire initialization and research phase.
Example Prompts
- "I think the last shell command broke my project structure. Please rollback the session to the checkpoint immediately preceding that action and restore the workspace files."
- "Show me the lineage of my current task. I want to see a tree view of all checkpoints so I can decide where to branch off for a new experimental fix."
- "Create a new child branch from checkpoint 'cp-842' to attempt a different dependency update strategy, but keep my current session running in the background."
Tips & Limitations
Always prioritize using the --json flag when chaining this skill with other automation scripts for machine-readable output. Remember that checkpoints are only created automatically before mutating tool calls; read-only operations will not trigger a checkpoint. Do not use ClawReverse for standard git operations; it is purpose-built for OpenClaw session state, not source control history. Finally, be aware that while rollback reverts the agent state, you must explicitly pass --restore-workspace if you intend to revert physical file changes in your environment.
Metadata
Not sure this is the right skill?
Describe what you want to build — we'll match you to the best skill from 16,000+ options.
Find the right skillPaste this into your clawhub.json to enable this plugin.
{
"plugins": {
"official-binhuangpjlab-clawreverse-skill": {
"enabled": true,
"auto_update": true
}
}
}Tags(AI)
Flags: file-write, file-read, code-execution