skills-management

The Skills Manager is a comprehensive orchestration tool designed for developers and power users of AI coding agents. It provides a unified workflow for discovering, maintaining, and organizing agentic skills across complex, multi-agent environments. Whether you are searching for new automation capabilities on the ecosystem marketplace or managing your local or project-scoped skills, this tool ensures your development agents remain performant and up-to-date. It simplifies the lifecycle of agent functionality by handling installation, version updates, quality reviews, and secure removals.

• Ecosystem Discovery: Search and browse the open agent skills ecosystem via the CLI, with direct access to marketplace resources and community-contributed tools.

• Scope Management: Easily move skills between user-level (global) and project-level scopes, ensuring configurations are shared appropriately across your team or kept personal.

• Multi-Agent Synchronization: Detect various coding agents on your system and synchronize skill availability, enabling seamless copying or moving of logic between Claude Code, Cursor, and other supported tools.

• Quality Assurance: Audit local skill implementations against best practices, including checks for naming conventions, description quality, security (XML frontmatter), and path integrity.

• Operational Control: Robust commands to list, detail, and remove skills safely, with built-in confirmation prompts to prevent accidental deletions of critical automation.

• Usage Note: Skills follow a hierarchy where user-scoped settings take precedence over project-level defaults. Always run reviews before deploying to shared projects to ensure best-practice compliance.

• Practical Tip: Use the 'python3 scripts/review_skill.py <name>' command frequently to maintain documentation standards and improve the trigger accuracy of your custom skills.

• Input/Output: Accepts standard bash commands via the CLI to interface with local directories; outputs include JSON summaries for programmatic integration or verbose text logs for human verification.

• Compatibility: Supports a broad range of agent environments; always verify agent-specific paths if you are performing cross-agent migrations.