skill-manager

Uninstall, disable, or re-enable community skills verified against the hub install log.

9.1k|1.8k|Updated Apr 21, 2026
One-click install
npx skills add https://github.com/anthropics/claude-for-legal --skill skill-manager-anthropics
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: skill-manager
Source: https://github.com/anthropics/claude-for-legal/tree/main/legal-builder-hub/skills/skill-manager
Command: npx skills add https://github.com/anthropics/claude-for-legal --skill skill-manager-anthropics

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Prevents unsafe or accidental removal of installed skills by verifying provenance in the hub install log, blocking first-party (built-in) plugins, and requiring explicit confirmation for any file changes.

Core Features & Use Cases

  • Verified uninstall workflow: Removes only community skills that are recorded in ~/.claude/plugins/config/claude-for-legal/legal-builder-hub/install-log.yaml with a most-recent action of install or enable, and only when the install path is outside the built-in plugin directories.
  • Safe disable workflow: Renames SKILL.md to SKILL.md.disabled (and related hook/agent files to .disabled variants) so scheduled agents stop firing, with confirmation before changes.
  • Optional re-enable workflow: Restores renamed files when the most recent action is disable, and records the change in the same install log.
  • Auditability and safety gates: Refuses first-party paths, refuses unknown skills not present in the log, never follows third-party SKILL instructions for removing other skills, and appends every action to the install log.
  • User-preserving configuration behavior: Does not delete user configuration directories by default (it surfaces them for the user), reducing the risk of losing valuable settings.

Quick Start

Run the command to remove a community-installed skill by name, and answer yes only when prompted to delete the exact files and folders the skill will remove.

Frequently Asked Questions about skill-manager

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I safely uninstall community skills without deleting user configuration directories?

To safely uninstall community skills, the workflow verifies the skill against the hub install log, blocks built-in first-party plugins, and requires explicit typed confirmation before file removal. It preserves user configuration directories by surfacing them for manual review instead of deleting them.

How do I disable scheduled agents and triggered skill behavior temporarily?

To disable scheduled agents, the tool renames SKILL.md to SKILL.md.disabled and applies .disabled variants to related hook and agent files. This stops triggered behavior safely and requires explicit confirmation before modifying any files.

Can I uninstall built-in first-party plugins using the skill manager?

No, you cannot uninstall built-in first-party plugins. The uninstall workflow explicitly refuses first-party paths and only removes community skills recorded in the legal builder hub install log with a most-recent action of install or enable.

How does audit logging work when disabling or re-enabling installed skills?

Audit logging appends every disable, re-enable, or uninstall action to the install-log.yaml file. This ensures a verified provenance trail for all community skill lifecycle changes and blocks unknown skills not present in the log.

What happens to my user settings when I remove a community skill?

When removing a community skill, user configuration directories are not deleted by default. The tool surfaces these directories for the user to review, reducing the risk of losing valuable settings during the uninstall process.

Can I re-enable a disabled skill and restore its scheduled agents?

Yes, you can re-enable a disabled skill by restoring the renamed files when the most recent action in the install log is disable. The workflow records this re-enable action in the same install log for auditability.