An MCP server for browsing CSES problems, submissions, and submitting solutions
Wellknown found it in public sources; nobody has proven control of it yet. Claiming takes one click if the repository is under your GitHub account, or a small file on your domain otherwise. Verified owners get the badge, 15-minute checks, status alerts, edits that outrank crawled data, and a ranking boost.
Agents can do it too: POST https://wellknown.network/api/v1/claims with {"agent":"qhung-cses-mcp","method":"well_known_file"} — machine-readable steps at claim.json, guide at /docs/claim.
Everything here was measured by our prober or read from a registry. Nothing is self-reported.
Attributed to the source that supplied each field. Treated as claims, not facts.
# CSES MCP Server *Your AI assistant's gateway to [CSES](https://cses.fi) (Code Submission Evaluation System)* 🧩 ## 🤔 What is this? `cses-mcp` is a Python-based MCP (Model Context Protocol) server that lets an MCP-compatible client (like Claude Desktop or Claude Code) act on your behalf on CSES: browsing problems, reading problem statements, submitting solutions, and checking submission history/verdicts. As far as I know, CSES has no public API, so all data access is via scraping authenticated HTML pages and posting to existing web forms using your own session cookie. ### Why? CSES is one of the best free problem sets for learning competitive programming, but working through it usually means a lot of manual tab- switching: reading a statement, writing code elsewhere, coming back to submit, then digging through a verdict page to figure out what went wrong. This project turns that whole loop into a conversation — ask an AI assistant to find you a problem, explain the statement, help you reason through an approach, submit your solution, and diagnose a wrong verdict, all without leaving the chat. The goal isn't to solve problems *for* you (though it can) — it's to make the feedback loop tight enough that you actually learn faster: get hints instead of full solutions when you want them, understand *why* a submission got Wrong Answer or Time Limit Exceeded, and use past-attempt history to see your own progress over time. --- ## 🚀 Quick Start Pick whichever fits — Option A if you just want to use it, Option B if you're editing the code. ### Option A — Copy-paste (no local clone) Paste this straight into your MCP client config — `uvx` fetches and runs the published package, no clone, no `uv sync`, no `.env` file to manage: ```json { "mcpServers": { "cses-mcp": { "command": "uvx", "args": ["qhung-cses-mcp@latest"], "env": { "PHPSESSID": "<your CSES session cookie>" } } } } ``` Requirements: - [`uv`](https://docs.as…
Mapped onto the structured taxonomy from declared text and observed tool names. Confidence shown for derived entries.
Every source is kept verbatim. Field changes are logged as events.