An AI mentor skill for DeepSeek Harness (and Codex / Claude Code): learn real system architecture by building it — stepped Chapters, hands-on Labs, behavior-test grading, layered hints, and AI Code Review, with a built-in DSH dashboard.
DSH integration
Native runtime
Author-claimed
Safety audit
Unaudited
Last verified
2026-08-29
License
Not specified by the author
01What can it help you accomplish?
Turn any open-source project into a hands-on, MIT-style engineering course
A stepped Chapter curriculum with lectures, Labs and behavior tests, plus AI grading (/promentor test) and AI Code Review
Developers and AI-coding-assistant users who want to learn a real system's architecture by building it
Learn a real system's architecture design instead of isolated algorithm puzzles
Layered, error-specific hints (/promentor hint) that point the way without giving the answer, and per-chapter progress and score tracking
Self-learners who prefer a dependency-ordered, hands-on path over reading source code at random
02How to install into DeepSeek Harness
Installation steps
- 01
Download the latest promentor.zip from Releases, unzip, then run `bash dsh-plugin/install.sh` from the unzipped promentor/ dir (recommended)
- 02
Or clone this repo, run `make build` first, then `cd /path/to/ProMentor && bash dsh-plugin/install.sh`
- 03
Alternatively, unzip and place `promentor/` under `.{YourAgent}/skills/`
Verify the integration
Not specified by the author
Rollback
- Uninstall with `bash dsh-plugin/uninstall.sh`
03DSH integration and capability boundaries
DSH Web GUI built-in plugin: a host data gateway (packages/host/promentor) + GUI panel (packages/client/ui-promentor) living in the deepseek-harness repo; this repo's dsh-plugin/ dir registers them via install.sh / uninstall.sh
Course generation (/promentor init)
a project directory opened in a supported AI coding assistant→course outline plus stepped Chapters (lectures, Labs, behavior tests)
writes course data under the project's `.promentor/` directoryGuided learning (/promentor learn)
a Chapter id (e.g. ch01)→lecture explanations, annotated source walkthrough, and prompts to hand-write the core logic
reads/writes the `.promentor/` course dataTesting & grading (/promentor test)
your implementation→a behavior-test report saying which tests passed, which failed, and why
Course dashboard
the `.promentor/` course data→completion, per-chapter status/score/attempts, and content-completeness warnings
DSH built-in panel reads the current session's working directory; standalone mode starts a local process that reads `.promentor/` at the project root
04Who is it for? When not to use it?
Good for
- Developers and AI-coding-assistant users who want to learn a real system's architecture by building it
- Self-learners who prefer a dependency-ordered, hands-on path over reading source code at random
05Compatibility, maintenance and safety notes
- The repo ships no build artifacts: the DSH dashboard's prebuilt bundle is distributed via the Release package, so source-mode installs must run `make build` first.
- The standalone dashboard fallback starts a global single-process local service (python3 scripts/serve.py) that reads `.promentor/` at the project root — a Python runtime is needed for that fallback path.
- The `/promentor` command only works inside AI coding assistants that support it — DSH (DeepSeek Harness), Codex, Claude Code, etc.; other environments have no built-in command.
TypeScript · actively maintained (latest release v0.3.0, 2026-08-13)
06Frequently asked questions
How do I install ProMentor on DeepSeek Harness?
Download the latest promentor.zip from Releases, unzip, then run `bash dsh-plugin/install.sh` from the promentor/ directory (recommended). Source-mode: clone the repo, run `make build`, then `bash dsh-plugin/install.sh`. The DSH dashboard is a built-in GUI plugin, so no separate local server is needed.
What does /promentor actually do?
It scans your project, generates a stepped Chapter course (lectures, Labs, behavior tests), guides you to hand-write the core logic, runs behavior tests and tells you what passed/failed and why, gives layered hints without spoiling the answer, and compares your implementation to the original source via AI Code Review.
Do I need to run a local server for the dashboard?
On DSH, no — the built-in panel reads the current session's `.promentor/` data directly with no local service. For Codex / Claude Code, the standalone fallback runs `python3 scripts/serve.py`, which starts a single global process that reads `.promentor/` at the project root.
How is ProMentor different from just reading source code?
It gives a dependency-ordered, stepped learning path instead of random jumping, makes you implement the core logic yourself rather than watching others code, and systematizes one system's design philosophy — with AI generating, explaining, grading and reviewing the course at zero content-production cost.
07Related DSH workflows
distilly
by titanwings
Meta-skill turning source material into a versioned Person Profile skill with incremental merge, correction layer, rollback, per-host install.
dsh-market
by dsh-market
插件市场 (plugin marketplace): 打开 Settings → Plugin Market 浏览/搜索/一键安装社区插件；安装源限制为 curated awesome-dsh-plugin 注册表（其余拒绝），默认禁用构建脚本（pnpm>=10 需显式开启）。
ai_animation
by unclecheng-li
Generate standalone animated HTML diagrams, presentations, notes, protocol visualizations, and UI demos from prompts.
aegis
by ganyuanran
Make AI coding agents architecture-aware: baseline-first, evidence-verified, drift-checked, and safe across long tasks.
08Data and sources
它的 Dashboard 是 **GUI 内置插件**
DSH 插件模式：host 数据网关（`packages/host/promentor`）+ GUI 面板
This page is generated from the project’s public documentation, repository metadata and a structured parse of DSH Plugins; last verified on 2026-08-29. Found an error? Submit a correction.
Best DeepSeek Harness Plugins
Twelve plugins worth installing first — picked from the whole catalog, across every category.