dsh-plugin-lion-skills
Run the following command in DeepSeek Harness:
dsh plugin install Lion-1209/dsh-plugin-lion-skills
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install Lion-1209/dsh-plugin-lion-skills in your terminal to install the plugin; the full source is available at https://github.com/Lion-1209/dsh-plugin-lion-skills .
About this plugin
Agents code fast, but they often skip the step you actually need: implementing before clarifying an ambiguous requirement, patching a symptom instead of chasing the root cause, or naming a variable with zero intention. dsh-plugin-lion-skills distills the 11 most common decision points in a developer workflow—requirement clarification, code review, debugging, naming, testing, task breakdown, pre-delivery verification, and more—into a structured Chinese skill suite that mounts on the DeepSeek Harness ctx.skills registry, so the agent loads the right mental framework automatically when the matching scenario triggers.
Each skill body is a plain Markdown file with name/description frontmatter, the same shape Claude Code uses, so you can edit and extend the suite without any build step. The suite registers as a bundled-rank 600 candidate, meaning your own project-level or user-level skills with the same name always take precedence. Malformed frontmatter fails loud at discovery rather than silently skipping a skill, and every skill declares resourceBase: directory so its body can reference sibling files such as evaluation fixtures, making it straightforward to keep refining trigger conditions and output quality over time.
If you already run DeepSeek Harness for day-to-day development and want your agent to follow more structured, layered Chinese guidance in scenarios like fuzzy requirements, unfamiliar codebases, design docs, and commit conventions, this suite works as a ready-made baseline that you can tune skill by skill to match your own project.
Use Cases
- Let the agent surface hidden assumptions before coding ambiguous requirements
- Guide the agent to trace root causes instead of patching symptoms when debugging
- Have the agent verify completeness and fix gaps before delivery
Best For
- Engineers using DeepSeek Harness for day-to-day development
- Teams that want structured Chinese agent guidance in their workflow
- Developers integrating the Lion-Skills methodology into their pipeline
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.