Session Continuity provides a robust checkpointing system that allows AI agents to save their exact state and resume unfinished work across multiple sessions without losing context.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install session-continuity
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install session-continuity using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Session Continuity?
Session Continuity is a specialized framework designed to solve the problem of context loss in AI agent workflows. Unlike standard chat history, this skill treats checkpoints as first-class, machine-readable artifacts that capture task-level resume signals, including specific progress, blockers, and the exact next action required. It ensures that even after a crash, manual reset, or multi-day break, your agent can start from the exact point of interruption rather than starting from scratch.
Built on the WAL (Write-Ahead Logging) Protocol, Session Continuity bridges the gap between active working memory and long-term task persistence. By integrating with existing session states, it proactively saves data before context limits are reached or when a session is closed, making it an essential tool for developers managing complex, multi-step engineering tasks through Openclaw Skills.
Session Continuity Use Cases
- Resuming a multi-day coding task where specific file states and pending decisions must be preserved.
- Recovering from session crashes or context window exhaustion without re-explaining the entire project.
- Handing off work between different agent instances or starting a fresh session to clear context clutter while keeping the goal intact.
- Explicitly marking milestones in a long-running project to create stable recovery points.
How Session Continuity Works
- The agent monitors the session for trigger signals such as high context usage (over 70%), close commands, or manual user requests.
- When saving, the agent reads the active SESSION-STATE.md and conversation history to gather specific progress, decisions, and blockers.
- A structured checkpoint file is written to the memory/checkpoints/ directory using a strict kebab-case naming convention.
- Upon starting a new session, the agent lists available checkpoints to identify unfinished tasks.
- When a resume command is issued, the agent verifies that all referenced file paths still exist to ensure the checkpoint isn't stale.
- The agent presents a recovery briefing and, upon user confirmation, executes the identified next action to continue the workflow.
Session Continuity Setup
To enable Session Continuity within your environment, ensure the directory structure exists and the skill is registered in your agent configuration.
# Create the required checkpoints directory
mkdir -p memory/checkpoints/
# Save your first checkpoint for a task
/checkpoint save build-auth-system --task "Implementing JWT rotation"
# List all active resume points
/checkpoint list
Session Continuity Data Schema & Taxonomy
The skill organizes data within a dedicated memory architecture to ensure durability and easy retrieval:
| File/Path | Purpose | Longevity |
|---|---|---|
memory/checkpoints/<name>.md |
Named, user-defined checkpoints for specific tasks. | Permanent until deleted |
memory/checkpoints/autosave.md |
Automatically generated state saved during close signals. | Overwritten by next auto-save |
SESSION-STATE.md |
The primary source of truth for active decisions and WAL logs. | Active Session |
memory/checkpoints/autosave-log.md |
A historical log of automated checkpoint activities. | Persistent |
Each checkpoint file contains structured sections for Task Summary, Progress, Next Action, Blockers, Key Decisions, and Relevant Context.
Session Continuity Advanced Features
- Deep Resume Stacking: Supports multi-day workflows by allowing checkpoints to survive beyond session death.
- Automated Context Monitoring: Internal protocol triggers an autosave when context window usage exceeds 70% to prevent information loss.
- Stale Path Detection: Automatically verifies file existence during the resume flow to warn users about outdated checkpoints.
- WAL Protocol Integration: Leverages Write-Ahead Logging to ensure that even per-message decisions are captured in the final checkpoint.
SKILL.md
Loading
Related Openclaw Skills
ValueScan Cryptocurrency Analysis
ValueScan is a sophisticated cryptocurrency analysis tool designed to track smart money flows, whale movements, and market signals for informed trading.
No Generator
A professional AI video editing skill that processes existing footage through a cloud GPU pipeline without generating new AI content.
OpenClaw SEO-GEO Workflow
A sophisticated automated runbook for managing high-volume SEO and GEO content production pipelines through specialized AI agents.
Geopolitics Analyst
A professional AI agent skill for generating neutral, structured, and data-driven analysis of modern international relations and geopolitical conflicts.
Content Polisher ZH
A specialized tool for humanizing AI-generated Chinese text and optimizing it for social media platforms like Xiaohongshu and WeChat.
Zhongshi Jianmo: Traditional Mantou Store Information
A specialized data skill for querying location, hours, and product details of the Zhongshi Traditional Mantou Store.
Featured*
