A standardized system for managing the lifecycle of long-term projects, ensuring reliable context recovery and cross-session persistence.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install openclaw-project-context-manager
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 openclaw-project-context-manager 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 Project Context Manager?
The Project Context Manager is a project-agnostic mechanism designed to handle multiple long-term projects within shared agent sessions. It solves the common problem of project context mixing with general conversation by implementing a file-based management system. This tool provides a set of Openclaw Skills that prioritize structured file records over volatile chat memory, ensuring that project progress, decisions, and states are preserved across session interruptions or archives.
By following a strict protocol for discovery, entry, and solidification, it allows developers to maintain a consistent project structure. It utilizes workspace-level rules and project-specific recovery indices to create a reliable boundary between general tasks and deep project work, making it indispensable for complex, multi-stage development workflows.
Project Context Manager Use Cases
- Managing multiple long-term development projects within a single shared AI session without context drift.
- Recovering detailed project states and progress milestones after a session has been archived or interrupted.
- Enforcing a standardized directory structure and metadata taxonomy across all user projects.
- Preventing sensitive project-specific data from being logged or leaked into general-purpose chat history.
How Project Context Manager Works
- The skill identifies the project environment by reading the workspace-level PROJECT_SYSTEM.md rules and the project registry.
- When a user requests to enter or resume a project, the manager scans the registry and directory for valid recovery entry files.
- Upon user confirmation, the skill executes the recovery protocol, reading index files, current progress reports, and key decision logs to restore the agent's context.
- During active work, the manager monitors for milestones to trigger 'solidification,' which saves checkpoints and updates the project's persistent state files.
- When exiting a project, the skill reinforces boundary rules to ensure subsequent messages are treated as general conversation and not appended to project records.
Project Context Manager Setup
To use this skill, ensure that python3 is installed in your environment. You can initialize a standard project structure by running the included initialization script:
python3 scripts/init_project.py '{
"projects_root": "/path/to/your/projects",
"project_name": "example-project",
"create_workspace_rule": true
}'
This will create the necessary directory structure and the registry entry required for discovery.
Project Context Manager Data Schema & Taxonomy
The Project Context Manager utilizes a structured file hierarchy to maintain state. Below is the primary data organization:
| Component | File/Path | Description |
|---|---|---|
| Workspace Rules | PROJECT_SYSTEM.md |
Global rules defining how projects are discovered and managed. |
| Registry | projects-registry.json |
A lightweight index mapping project names to their absolute filesystem paths. |
| Recovery Entry | 00_Recovery_Entry.md |
The primary file used to bootstrap project context upon entry. |
| Progress Index | 00_Index_and_Progress.md |
A summary of completed work, current tasks, and breakpoints. |
| Decision Log | 99_Key_Decisions.md |
A persistent record of architectural and technical choices. |
| Checkpoints | checkpoints/ |
A directory containing snapshots of the project at various stages. |
Project Context Manager Advanced Features
- Automated project discovery via a centralized registry in the Openclaw Skills workspace.
- Milestone-based state solidification that prompts users to save progress after significant file changes or decisions.
- Support for custom checkpointing logic and session backups for high-stakes development.
- Flexible initialization parameters allowing for custom root directories and rule filenames.
- Explicit entry/exit protocols to maintain context integrity in multi-agent or multi-project environments.
SKILL.md
Loading
Related Openclaw Skills
Treasury Scenarios Skill
A specialized AI agent skill designed to query business scenarios, interface combinations, and visual workflows for the Postal Savings Bank of China (PSBC) treasury system.
Adaptive Team Research
An adaptive multi-agent orchestration skill that automates software project reviews through structured collaboration modes and a rigorous three-round consensus workflow.
Resume Evaluator
A professional AI-driven recruitment tool that generates structured resume evaluation reports from an HR expert perspective.
ClawFight Arena
ClawFight Arena is a competitive AI agent battle platform where users register virtual lobsters to compete in quiz challenges for ELO rankings.
Mathbox
A local, high-speed reference tool providing immediate access to Mathbox implementation patterns, debugging guides, and performance best practices.
Fontpreview
A fast, local-first documentation utility for Fontpreview patterns and developer best practices.
Featured*
