szw321127

szw321127 / llm-wiki-for-code

Public

为llm服务的代码片段知识库

17
1
100% credibility
Found May 07, 2026 at 17 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
JavaScript
AI Summary

A tool that builds a local knowledge base and visual graph in code projects to help AI coding assistants recall practices, decisions, and evidence for consistent development workflows.

How It Works

1
📰 Discover the tool

You hear about Project Knowledge, a handy way to help your AI coding buddy remember your project's favorite ways of doing things so it doesn't repeat old work.

2
📥 Add to your project

You download the files and place them in your coding project's folder to get started.

3
📁 Create knowledge folder

You run a simple setup command to make a special folder in your project that stores tips, choices, and lessons learned.

4
💡 Check for smart tips

Before starting a new coding task, you ask the tool for relevant past practices and top suggestions to guide your work smoothly.

5
💻 Code with AI help

Your AI assistant uses the tips to give consistent advice, making your coding faster and matching your project's style.

6
📝 Save what you learned

After the task, you quickly record the key decisions, changes, and what worked well for future use.

7
🗺️ View knowledge map

You open a colorful interactive map showing all your project's connected ideas, options, and rules at a glance.

🎉 AI remembers forever

Now every coding session starts with your project's wisdom, keeping everything consistent and saving tons of time.

Sign up to see the full architecture

6 more

Sign Up Free

Star Growth

See how this repo grew from 17 to 17 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is llm-wiki-for-code?

This JavaScript tool builds a persistent llm code wiki in your project's `.project-knowledge/` directory, storing practices, options, rules, and session decisions as linked Markdown files with a knowledge graph. It integrates with LLM GitHub Copilot, Claude Code, or Codex via plugins, letting you run `pk:preflight` before tasks for relevant recommendations and `pk:auto-crystallize` afterward to capture outcomes. Solves the pain of LLMs rescanning code repeatedly or losing decisions to chat history in long-running codebases.

Why is it gaining traction?

Unlike static AGENTS.md files, it tracks evolving practices with adoption counts, evidence previews from stable paths, and top-3 recommendation pools, plus a local graph server (`pk:serve`) for browsing relationships. LLM wiki for Claude code or Codex stands out with linting, governance commands like `pk:govern`, and Obsidian compatibility, keeping context lean without opaque vector stores. Developers hook into it via CLI or AI skills for llm github integration that feels project-owned.

Who should use this?

Backend or fullstack devs on mature JavaScript repos using LLM GitHub Copilot or Claude for tasks like HTTP clients or schedulers, tired of inconsistent implementations across sessions. Teams maintaining llm github projects where agent outputs need reviewable evidence trails, not buried in transcripts. Suited for anyone running llm github actions or courses needing durable codebase conventions.

Verdict

Try it if you're deep into Codex or Claude Code workflows—solid docs, tests, and plugin setup make onboarding fast despite 17 stars and 1.0% credibility score signaling early maturity. Pairs well with llm github repository tools but watch for scaling in huge codebases.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.