muse

What This Skill Does

Muse is a sophisticated orchestrator for the OpenClaw AI agent ecosystem, designed to bridge the gap between individual coding sessions and collective team knowledge. It functions as a historical data engine, indexing your entire coding history, session transcripts, and project-specific documentation. By connecting these disparate data points, Muse enables ClawBot to evolve from a basic autocomplete tool into an autonomous agent capable of cross-referencing past architectural decisions, mediating team technical discussions, and executing complex, multi-step development tasks across your codebase. It serves as both a memory bank and a command center for autonomous agent orchestration.

Installation

To integrate Muse into your environment, use the OpenClaw CLI tool: clawhub install openclaw/skills/skills/alexander-morris/muse. After installation, ensure you have authenticated by running tribe login. Note that advanced agent orchestration features require the -beta flag. Ensure your terminal environment supports TUI components if you intend to use the monitoring or dashboarding features.

Use Cases

1. Legacy Knowledge Retrieval: Quickly query how a specific authentication module was implemented six months ago across multiple sessions.
2. Autonomous Feature Design: Spawn sub-agents to handle specific tasks (e.g., refactoring or testing) while maintaining context from your primary project documentation.
3. Team Synchronization: Use the Knowledge Base (KB) functionality to sync team-wide deployment patterns and coding standards, ensuring all agents adhere to project-specific constraints.
4. Incident Response: Analyze past session events and error logs stored within Muse to rapidly diagnose recurring bugs or performance regressions.

Example Prompts

1. "Muse, search through my last 30 days of sessions to find the implementation details for the OAuth2 middleware and create a summary for the team."
2. "Spawn a subagent named 'bug-squasher' to analyze the current login issues, apply the fix using the pattern we established in last Tuesday's session, and output the results for my review."
3. "Save these documentation notes to the knowledge base: 'All API endpoints must now include the x-request-id header for distributed tracing'."

Tips & Limitations

• TUI Requirements: Commands like monitor and dashboard must be run in an interactive terminal session and are not suitable for CI/CD pipeline automation.
• Authentication: Keep your sessions active with tribe login. If the agent loses context, perform a tribe kb sync to ensure local knowledge is updated with the latest remote data.
• Beta Caution: The orchestration commands (MUSE and CIRCUIT) are currently in beta; ensure you have a clean git state before spawning autonomous subagents to prevent unintended changes to your production codebase.