By technomensch
Capture, retrieve, and synchronize structured project knowledge—architecture decisions, lessons learned, and session summaries—into a local knowledge graph with git metadata, FTS5 search, and GitHub issue/PR linking.
Add a new category to an existing knowledge graph with optional custom prefix
Document lessons learned, problems solved, and patterns with git metadata tracking
Scan active knowledge graph for potentially sensitive information before public sharing
Interactive wizard to set up pre-commit hooks for sensitive data detection
Create Architecture Decision Records with auto-filled git metadata, sequential numbering, and index auto-update
Creates Architecture Decision Records — interactive wizard with auto-numbered files, git metadata, template population, and index updates. Uses kg_capture MCP tool for platform-agnostic writes.
**Role:** Parse large chat history files, lesson documents, and session logs to extract structured insights and relationships for the knowledge graph. Prevents the main context window from being consumed by 2000+ line source files. Also handles the full KG entry extraction pipeline when delegated from `update-graph`.
Reviews knowledge graph entries for quality, completeness, and proper bidirectional linking. Use when creating or updating KG entries to ensure they meet quality standards.
Captures a single lesson learned from the current session — problem, solution, context, and git metadata. Uses kg_capture MCP tool for platform-agnostic writes.
Detects IDE environment and auto-configures MCP server registration when kg_* tools fail. Enables knowledge graph operations across platforms without manual setup.
Auto-surface ADR creation when user makes architectural decisions or chooses between technical approaches.
Consult the knowledge graph before any recommendation is made during brainstorming.
Route "capture that" / "remember that" requests to the correct destination (memory, lesson, or ADR) via auto-detection.
Route explicit doc-update requests to the correct command, ensuring standards validation and changelog enforcement.
Use when the user is about to push to origin, open a PR, create a PR, merge changes, or says "push and merge", "push and merge with admin", "push to origin", "finishing up", or "ready to push". Scans for all user-facing docs affected by the current branch's changes before any update wizard runs.
Admin access level
Server config contains admin-level keywords
Executes bash commands
Hook triggers when Bash tool is used
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Modifies files
Hook triggers on file write and edit operations
Modifies files
Hook triggers on file write and edit operations
Uses power tools
Uses Bash, Write, or Edit tools
Uses power tools
Uses Bash, Write, or Edit tools
Structured knowledge capture, lesson-learned documentation, and cross-session memory for Claude Code projects.
Version: 0.5.11 Status: Actively developed and in daily use
Documentation: https://kmgraph.stayinginsync.info
Buy me a coffee if you find this useful - https://buymeacoffee.com/technomensch
This is a platform-agnostic knowledge graph that was developed entirely using Gemini and Claude, leveraging very specific context and detailed natural language prompting.
It is designed to take chat sessions with large language models (LLMs) and turn them into a searchable, institutional knowledge library.
The cool thing is, it helps users grab the important stuff (lessons learned, architecture decisions, recurring patterns, etc...) inside the development workflow without having to stop chatting.
Then, users can easily look up that information not only in their current chat, but also in any other chat session, even if they switch to a totally different LLM!
The key lies in the simple approach of embedding the knowledge directly within the project itself. This ensures the knowledge is always immediately available whenever and wherever the project is opened. Should the library become excessively large, users have the option to transfer it to an external third-party via MCP servers.
A Claude Code plugin that provides:
Paste INSTALL.md into any AI assistant for automated setup on any platform — Claude Code, Codex CLI, Cursor, Windsurf, Continue.dev, JetBrains, VS Code, Aider, or local LLMs.
Claude Code users: Run claude plugin install kmgraph or load with claude --plugin-dir /path/to/knowledge-graph, then run /kmgraph:init.
Codex CLI users: Run codex plugin marketplace add technomensch/knowledge-graph then codex plugin add kmgraph@knowledge-management-graph.
See the Quickstart for prerequisites and troubleshooting.
Pull the latest version and run /kmgraph:init in any project that uses it. The upgrade wizard checks what has changed, previews any updates to existing files, and asks for confirmation before writing, or changing, anything. Existing knowledge graph content is never overwritten.
Quick Reference: See CHEAT-SHEET.md for one-page quick reference guide Detailed Guide: See Command Guide for comprehensive command documentation with learning paths
/kmgraph:init — Initialize new knowledge graph with wizard-based setup/kmgraph:capture-lesson — Document lessons learned with git metadata tracking/kmgraph:create-adr — Create an Architecture Decision Record with automatic implementation commit capture/kmgraph:status — View active knowledge graph info and quick reference/kmgraph:recall — Search across all memory systems (lessons, decisions, knowledge)/kmgraph:update-graph — Extract knowledge graph entries from lessons/kmgraph:add-category — Add a new category to existing knowledge graph/kmgraph:session-summary — Create summary of current chat session/kmgraph:list — Display all configured knowledge graphs/kmgraph:switch — Change active knowledge graph/kmgraph:check-sensitive — Scan knowledge graph for potentially sensitive information/kmgraph:config-sanitization — Interactive wizard for pre-commit hook setup/kmgraph:extract-chat — Extract chat history from Claude, Gemini, and Codex CLI logs (--source codex for Codex sessions)/kmgraph:update-doc — Update plugin/project documentation (--user-facing) or KG content/kmgraph:init-personal-kg — Initialize a personal knowledge graph at ~/.kmgraph/ shared across all projectsnpx claudepluginhub technomensch/knowledge-graph --plugin kmgraphMake your AI agent code with your project's architecture, rules, and decisions.
Claude + Obsidian knowledge companion. Sets up a persistent, compounding wiki vault (Karpathy's LLM Wiki pattern). v1.7 "Compound Vault" + v1.8 methodology modes close 5 of 5 priority gaps from the May 2026 compass artifact. Ships: substrate alignment with kepano/obsidian-skills, default Obsidian CLI transport, hybrid retrieval (contextual prefix + BM25 + cosine rerank per Anthropic's Sept 2024 research), per-file advisory locking for multi-writer safety, pre-commit verifier agent, AND methodology modes (LYT / PARA / Zettelkasten / Generic) for first-class organizational support no other Claude+Obsidian competitor offers. v1.7.x audit closure: every BLOCKER + HIGH + MEDIUM + LOW finding from the v1.7.0 audit is CLOSED or DEFERRED-with-rationale. Optional DragonScale Memory extension (log folds, deterministic addresses, semantic tiling lint, boundary-first autoresearch).
Upstash Context7 MCP server for up-to-date documentation lookup. Pull version-specific documentation and code examples directly from source repositories into your LLM context.
Complete AI coding workflow system. Self-correcting memory + persistent FTS5-indexed research wikis + auto-research loop + multi-LLM council on a single SQLite store. 33 skills, 8 agents, 22 commands, 37 hook scripts across 24 events. Cross-agent via SkillKit.
AI-powered wiki generator for code repositories. Generates comprehensive, Mermaid-rich documentation with dark-mode VitePress sites, onboarding guides, deep research, and source citations. Inspired by OpenDeepWiki and deepwiki-open.
Comprehensive C4 architecture documentation workflow with bottom-up code analysis, component synthesis, container mapping, and context diagram generation