hackyeaster-solver
MCP server enabling Claude to solve HackyEaster CTF challenges through web scraping, decoding, image analysis, and progress tracking.
README
HackyEaster Solver
MCP server for Claude-driven HackyEaster CTF solving. Uses Scrapling for page fetching and exposes CTF utility tools so Claude can reason through challenges, submit answers via Playwright, and iterate.
Setup
Quick install (no clone needed)
Add to your ~/.mcp.json:
{
"mcpServers": {
"hackyeaster": {
"command": "uvx",
"args": ["--from", "git+https://github.com/cloudwarriors-ai/hackyeaster-solver", "hackyeaster-mcp"]
}
}
}
Restart Claude Code. The he_* tools will appear in the MCP servers panel.
Local development
If you want to modify the server or keep persistent state in a known location:
git clone https://github.com/cloudwarriors-ai/hackyeaster-solver.git
cd hackyeaster-solver
uv sync
Then in ~/.mcp.json, point at your local clone:
{
"mcpServers": {
"hackyeaster": {
"command": "bash",
"args": ["-c", "uv run --directory /path/to/hackyeaster-solver python -m hackyeaster_mcp.server"]
}
}
}
Tools
Scraping
| Tool | Description |
|---|---|
he_discover_challenges |
Scrape hackyeaster.com and return available challenges |
he_fetch_challenge |
Fetch and parse a specific challenge page |
he_fetch_raw |
Fetch any URL and return raw HTML |
CTF Utilities
| Tool | Description |
|---|---|
he_decode |
Decode: base64, hex, url, binary, morse, decimal, octal, base32 |
he_transform |
Transform: caesar, rot13, swap_pairs, reverse, rail_fence, vigenere, xor, atbash |
he_analyze_image |
Download image for visual analysis + QR/barcode detection |
he_hash |
Compute hashes: md5, sha1, sha256, sha512 |
State
| Tool | Description |
|---|---|
he_log_attempt |
Record a solve attempt |
he_get_progress |
View all attempts and solved challenges |
he_validate_flag |
Check if text matches flag format he20XX{...} |
How It Works
Claude is the solver. The MCP server provides the tools; Claude provides the reasoning.
he_discover_challenges— scrape the site for challengeshe_fetch_challenge— read challenge content, hints, embedded data- Claude reasons about the puzzle, calls
he_decode/he_transform/he_analyze_imageas needed - Playwright MCP (separate, already available in Claude Code) handles form interaction and answer submission
he_log_attempttracks what was tried and what worked
Requirements
- Python 3.11+
- uv
- Claude Code with MCP support
- Playwright MCP server (for answer submission when challenges go live)
Architecture
Claude (solver brain)
|
+-- hackyeaster MCP -- Scrapling fetch + CTF utils + state
|
+-- playwright MCP --- form interaction, submit answers, screenshots
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。