Keen Code
Context-efficient CLI coding agent built in Go — minimalist terminal AI with turn memory and lazy-loaded skills.
Your Citability Score
62/100- Evidence: Add: At least one evidence link, Multiple evidence sources, Demo URL.
- Trust: Add: Contact information, Support URL, Privacy policy, Terms of service.
- Freshness: Add: Has version number, Launch date provided, Reviewed within 90 days.
Promotion (Boost) does not change this score — it only changes ordering. This number reflects real, verifiable citability.
Frequently asked questions
- What is Keen Code?
- Context-efficient CLI coding agent built in Go — minimalist terminal AI with turn memory and lazy-loaded skills.
- What does Keen Code do?
- Keen Code is an open-source, terminal-based AI coding agent written in Go. It focuses on context efficiency through a "turn memory" system that discards raw tool traces after each turn, retaining only distilled summaries of files changed and failed commands. Unlike agents that load large MCP…
- Who is Keen Code for?
- Software engineers and developers who prefer terminal-native workflows and want a lightweight, context-efficient alternative to heavier coding agents.
- Is Keen Code verified?
- Keen Code is listed on CitableHub with a citability score of 62/100, computed from verifiable profile evidence.
Keen Code
InvitedContext-efficient CLI coding agent built in Go — minimalist terminal AI with turn memory and lazy-loaded skills.
Structured for AI systems to extract and cite.
Citability Score
Citable Outcome
Ship code faster from the terminal with an AI agent that uses minimal context tokens — keeping costs low and responses focused.
About
Keen Code is an open-source, terminal-based AI coding agent written in Go. It focuses on context efficiency through a "turn memory" system that discards raw tool traces after each turn, retaining only distilled summaries of files changed and failed commands. Unlike agents that load large MCP schemas upfront, Keen Code uses lazy-loaded markdown-based Skills, retrieving JSON schemas only at runtime. Ships with 6 core tools (read_file, write_file, edit_file, glob, grep, bash) and supports multiple LLM providers including Anthropic, Claude, OpenAI, Gemini, and DeepSeek.
What makes it different
- Turn memory system discards raw traces — retains only distilled summaries for minimal token usage
- Lazy-loaded markdown Skills instead of upfront MCP schema loading
- Built entirely by AI agents under human orchestration — full development history preserved
- Multi-provider support: Anthropic, OpenAI, Gemini, DeepSeek, Kimi and more
Tags & Classification
Links & Transparency
Cite this Project
@misc{citablehub_keen-code,
title = {Keen Code},
url = {https://citablehub.com/p/keen-code},
note = {Listed September 10, 2026. CitableHub ID: CH-VER-676417},
year = {2026}
}Keen Code. (2026). CitableHub Software Index. https://citablehub.com/p/keen-code.
"Keen Code." CitableHub, 2026, https://citablehub.com/p/keen-code.
