skill-bartender
Run the following command in DeepSeek Harness:
dsh plugin install MJorgin/skill-bartender
Paste the following prompt into your AI chat to install this plugin:
To install this plugin in DeepSeek Harness, please execute dsh plugin install MJorgin/skill-bartender. Source: https://github.com/MJorgin/skill-bartender
About this plugin
In DeepSeek Harness, your agent often struggles with skill overload, loading too many unnecessary skills or even degrading conversation quality with incorrect ones. skill-bartender is designed to solve this "buffet-style" skill management problem, ensuring your agent only loads skills when absolutely necessary and preventing invalid instructions from clogging conversation history. The plugin's core strength lies in its "Laziness Ladder" strategy and "Safe Cellar" mechanism. It loads skills in a strict priority order—from simple tools to complex workflows—and refuses to install anything when uncertain. Furthermore, any new skill installation goes through a mandatory isolation, scanning, and human approval process, completely eliminating silent installs and security risks. This plugin is ideal for developers who need precise control over skill invocation, especially those building complex Agent workflows. Whether you require custom routing rules or high security standards, skill-bartender uses its intelligent routing and auditing features to help you find the "right cocktail" for every task within a vast ecosystem of skills.
Screenshots
Use Cases
- When the skill catalog is too large, slowing down agent responses
- When strict security policies are needed to prevent malicious skill injection
- When you want to define intelligent routing rules for specific task types
Best For
- Developers building complex Agent workflows
- Enterprise users concerned about prompt injection security
- DSH plugin users needing fine-grained control over skill loading
Related Plugins
A method pack that makes AI coding agents plan against your real baseline, prove completion with fresh evidence, and reduce reworks and unsafe changes.
Turns the DeepSeek Harness session into a captain that builds a durable sub-agent team, splits goals into dependency-aware tasks, and coordinates work via direct messages and a live Web UI.
Gives coding agents design judgment, letting Claude Code, Cursor, and 70+ agents generate and iterate high-quality UI, presentations, and graphics right from the terminal.
Run the Pi ecosystem's plugins on DeepSeek Harness, unmodified, via a compatibility layer that implements Pi's public extension ABI on DSH's native services.
