#126126th of 143 in Coding agents and AI editors
corecoder
Minimal Python coding agent written to be read and forked
- Stars
- 1.8k
- License
- MIT
- Last commit
- Oct 2026
- Last release
- Oct 2026
- Language
- Python
Missed the latest GitHub check; these facts may be older.
Overview
CoreCoder is a working terminal coding agent in about 2,816 lines of Python (1,322 in the engine). It runs a tool loop with eight tools, three-tier context compaction, sub-agents, MCP stdio tools, plan mode, hooks and /undo. It is meant as a readable reference to study and fork, backed by eight source-reading essays, not as a daily driver.
Who it is for: Engineers who want to learn how coding agents work and fork their own
Strengths
- Whole engine is small enough to read in one sitting, with a file-by-file code map
- Works with any OpenAI-compatible endpoint; optional LiteLLM backend for other providers
- Mutating tools ask for consent; -p mode refuses them unless --yes is passed
- Eight bilingual essays walk through the actual code; 227 tests
Weaknesses
- Bash safety is a regex blacklist, not a sandbox; unsafe for untrusted input
- No fallback model chain or hard dollar budget; retry is plain exponential backoff
- Sub-agents run synchronously; no RAG; MCP client supports tools only
- Author states it is not intended as a daily driver
What it needs
- no GPU
- Needs Python 3.10+, LLM API key (OpenAI-compatible)
- Models: OpenAI, DeepSeek, Ollama, OmniRoute, Kimi
Also in Coding agents and AI editors
See all 143| Rank | Project | Score |
|---|---|---|
| 1 | ArchonYAML workflow engine that runs AI coding agents through fixed dev steps | 84 out of 100 |
| 2 | claw-codeRust CLI coding agent harness using Anthropic and OpenAI-compatible APIs | 83 out of 100 |
| 3 | Open SWEAgent that turns issues into pull requests, reviews PRs and monitors CI | 83 out of 100 |
| #4 | piTerminal coding agent with extensions, skills, and an SDK | 80 out of 100 |
| #5 | oh-my-openagentTerminal coding agent with multi-model orchestration and git-backed memory | 80 out of 100 |