skill-DSA — A Visual Skill for Data Structures and Algorithms

Platform Compatibility — Test Results (2026-07-27)

Platform	Status	Notes
Claude	Works	Runs well both by reading the repository link and as context inside a Project.
Grok	Works	Performs almost as well as Claude.
Gemini	Partial	Cannot use the GitHub URL. After downloading the repo as a ZIP, it only works through a Gem. Cannot render the interactive diagrams.
GitHub Copilot	Partial	Must read the specific folder / SKILL.md file to run the skill; cannot run it from the GitHub project URL.
Microsoft Copilot	Not working	Cannot read the GitHub URL to run the skill; .zip files are not supported either.
ChatGPT	Not working	Cannot run the skill.

---

A cross-platform AI teaching skill: grounded in 74 Java reference implementations as its source of truth, it delivers a full course flow — concept explanation → interactive HTML diagram → step-by-step reasoning → close reading of the reference code → cross-referenced walkthrough → summary → prerequisite/follow-up recommendations — and you can ask questions at any point.

Covers 44 topics: 18 data structures (sparse arrays, queues, stacks, linked lists, recursion, hash tables, the binary tree family, and more) and 26 algorithms (the eight classic sorts, four search methods, KMP, greedy algorithms, Prim/Kruskal/Dijkstra, dynamic programming, Huffman coding, and more).

---

Source of the Java Code

The reasoning and teaching structure behind every original Java file stored in this skill's assets/ directory come from the《数据结构与算法》(Data Structures and Algorithms) course by 尚硅谷韩顺平老师.

Building on that foundation, this project reworked the code: split into packages by difficulty (B / Bp / N / ZA), unified to JDK 11 style, with a number of defects fixed that only surfaced under specific conditions, and teaching comments ("concept + step-by-step reasoning") filled in throughout.

---

Quick Start

On Claude (native command)

1. Place the entire skill-DSA/ directory into Claude Code's skills directory:
   • Personal: ~/.claude/skills/skill-DSA/
   • Project: <project root>/.claude/skills/skill-DSA/
2. Type /skill-DSA in the conversation to start (if your version normalizes commands to lowercase, use /skill-dsa).
3. On claude.ai: upload this directory to a Project's knowledge base, then type /skill-DSA in the conversation.

On Claude, interactive diagrams render directly as HTML Artifacts (if rendering fails in the mobile app, the skill has a built-in fallback: "save as .html and open in a browser").

On Other Generative AI Platforms (ChatGPT / Gemini / Grok, etc.)

These platforms have no skill mechanism, so you start it by loading the prompt and attaching the code files:

1. Load the master prompt: open GeneralSkill_DSA_Prompt.md and paste it in full at the start of the conversation; or, on platforms that support file upload (ChatGPT Projects / custom GPTs / Gemini Gems), add the file to the knowledge base.
2. (Recommended) Attach the reference code: upload the .java files from assets/src/ as well. It runs without them, but any code shown will be marked "not reference code."
3. Type /skill-DSA to start, then follow the prompts to pick your natural language and programming language.
4. Type /skill-DSA-end when you're done studying.

Note: the free tier of ChatGPT, Gemini, and similar platforms may not be able to render interactive diagrams directly. The skill states this automatically at startup, outputs the diagram as complete HTML code, and includes three-step instructions for saving it as .html and opening it in a browser.

Via a Public GitHub Repository

Once this directory is published as a public repository, anyone can:

• git clone it or download the ZIP, then load it using either method above;
• On AI platforms that can read URLs, give the model the raw link to GeneralSkill_DSA_Prompt.md and then type /skill-DSA.

---

Common Commands

Command	What it does
/skill-DSA	Start the skill (pick languages → show catalog → pick a topic and begin)
continue	Move to the next stage at a checkpoint
Just type your question	Ask anything at any time (returns to the course progress anchor when answered)
/skill-DSA-end	Exit and output the course report

Directory Structure

skill-DSA/
├── SKILL.md                      # Native Claude command entry point
├── GeneralSkill_DSA_Prompt.md    # Cross-platform master prompt (single source of all rules)
├── README.md                     # This file
└── assets/src/
    ├── a_DataStructures/         # Data structure reference code (18 topics, difficulty prefixes B/Bp/N/ZA)
    └── b_Algorithm/              # Algorithm reference code (26 topics, difficulty prefixes B/Bp/N/ZA)

Assumed Prior Knowledge

The skill assumes you already have basic computing knowledge (variables, loops, functions).

If you're starting from zero, watch Harvard's CS50x first: https://www.youtube.com/playlist?list=PLhQjrBD2T380hlTqAU8HfvVepCcjCqTg6