dsh-xingtu-skills
Run the following command in DeepSeek Harness:
dsh plugin install xingtu1996/dsh-xingtu-skills
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install xingtu1996/dsh-xingtu-skills in your DeepSeek Harness terminal to install the plugin; the source repository is available at https://github.com/xingtu1996/dsh-xingtu-skills .
About this plugin
AI agent skill files are typically copied between tools by hand, breaking the moment your stack changes. This plugin packages 26 production-grade skills into one standard DSH bundle. After install, the session skill catalog loads everything automatically and triggers each skill on demand—no manual copying, no per-skill toggle.
Three series cover the full loop: Caveman (13 skills) handles token compression and context management so prompts stay lean; Ponytail (6 skills) pushes minimalism and debt awareness so the agent asks why before it builds; Engineering Practice (7 skills) walks through investigate-first, lean-build, safe-refactor, surgical-patch, and verify-and-stop. All skills follow the cross-tool name, description, when_to_use standard and also work with Claude Code, CodeBuddy, Codex, Cursor, and Gemini CLI.
Zero runtime dependencies, no bundled copies of third-party code, and both static and clean-boot validation are built in. It is aimed at developers and small teams who want a single, consistent skill layer that raises token efficiency, code quality, and engineering discipline in one go, on top of whatever AI coding harness they already use.
Use Cases
- Trigger skills on demand during AI coding sessions to compress context and control token spend
- Enforce a disciplined engineering loop: investigate first, lean build, safe refactor, verify and stop
- Unify skill configuration across Claude Code, Cursor, Gemini CLI, and other tools without file-by-file copying
Best For
- Developers working with DeepSeek Harness or any DSH-compatible plugin host
- Teams looking to cut AI coding token costs while keeping prompts efficient
- Engineers who value code quality and engineering discipline over maximal generation
Related Plugins
A unified suite combining hot runtime injection, task-aware thinking-mode routing, and a graded session protocol with red-team gates to sustain model diligence across long-horizon inference.
ModLens is a vision plugin for DeepSeek Harness that gives text-only models sight by reading images pasted directly into chat, with zero-config setup and multiple vision engines.
On-demand vision for text-only DeepSeek Harness agents: built-in free keyless vision chain and 14 vision tools, routing image turns as tool calls to vision models with pixel fidelity, no Python needed, one-command install.
Give text-only models in DeepSeek Harness eyes, enabling image Q&A, long-screenshot OCR, UI restoration, and GUI visual tasks.