VividKit Crystal clear AI coding
VI
Choose Your Setup Get Started with AgentKit Continue with ClaudeKit Migrate from ClaudeKit to AgentKit Clean cutover from ClaudeKit What Is AgentKit? How AgentKit Works AgentKit Kits, Skills, Agents & Hooks AgentKit Runtime Support AgentKit Desktop App AgentKit Configuration AgentKit CLI Quick Reference AgentKit Skills Cheatsheet AgentKit Workflows AgentKit Statusline Update AgentKit CLI & Kits AgentKit Helper Clean reinstall of AgentKit AgentKit Troubleshooting What is ClaudeKit? ClaudeKit CLI Guide Command Finder Commands Cheat Sheet Workflow Recipes CLI Quick Reference Claude Mechanics Permission Modes ClaudeKit Hooks Coexistence ClaudeKit x Codex CK → Other coding tools How ClaudeKit Works Inside ClaudeKit Finding Your Unknowns Session Recovery Fix from Logs UI/UX Guide Accounts Switcher IDE Config Remote Control Codex App System Diagramming UI Review Gate Promotions Support VividKit

Quick Ref / Command

Engineer Kit

/ak:graphify

Graphify Knowledge Graph Builder

Turn code, docs, papers, images, audio, or video into a queryable graph with local AST extraction, local transcription, semantic extraction, reports, and MCP querying.

01

Pick target

02

Install correctly

03

Build graph

04

Read outputs

Rule 01

Use graphs to navigate relationships, not to replace source-level inspection.

Rule 02

Start with the report and visualization, then query graph.json for targeted paths.

Rule 03

Separate extracted facts from inferred or ambiguous relationships.

Execution Map

Core lanes from input to output, aligned with the command's real execution path.

01

Start

Input, scope, route

  1. 1 Pick target Choose a path when architecture, cross-file relationships, god nodes, or token-efficient navigation are needed.
  2. 2 Install correctly Use the PyPI package graphifyy if the runtime is missing; do not confuse it with unrelated graphify packages.
02

Work

Agent / skill execution

  1. 3 Build graph Run graphify . or graphify /path/to/project to produce graphify-out artifacts.
  2. 4 Read outputs Inspect graph.html for visualization, GRAPH_REPORT.md for god nodes and suggested questions, graph.json for persistent queries, and cache/ for incremental state.
03

Verify

Gate, review, validation

  1. 5 Query via MCP When MCP mode is needed, serve graphify-out/graph.json with python -m graphify.serve and use query_graph, get_node, get_neighbors, or shortest_path.
  2. 6 Combine with scout Use Graphify for high-level structure and /ak:scout for specific file discovery.
04

Close

Report, handoff, artifact

  1. 7 Update incrementally Rerun after changes; SHA256 cache means only changed files are reprocessed.
  2. 8 Name privacy boundary Report that code AST and audio/video transcription are local, while docs/images use the configured model provider for semantic extraction.

Syntax, arguments, subcommands

How to invoke the skill: syntax, positional arguments, shared options, then each subcommand with its own syntax and outcome.

Syntax

/ak:graphify [path] [--mcp|--report|--watch]

Arguments

[path]

Input path

Folder or bounded document set to graph. Choose the smallest useful scope; the command writes graphify-out/ relative to the run location and does not automatically include unrelated secrets, build output, or archives.

/ak:graphify ./src --report

Shared options

--report

Report emphasis

Center the result on GRAPH_REPORT.md findings such as prominent nodes, surprising relationships, and suggested questions. Report claims still need source review.

/ak:graphify ./src --report
--watch

Watch rebuilds

Rebuild incrementally as files change. This starts ongoing observation and repeated writes; it does not authorize a background process beyond the requested run.

/ak:graphify . --watch
--mcp

MCP graph queries

Prepare graphify-out/graph.json for MCP-backed query tools such as query_graph, get_node, get_neighbors, and shortest_path. Server installation and client configuration remain separate requirements.

/ak:graphify . --mcp

Sample Prompt

Concrete invocations for each flag, subcommand, mode, or route available in the live workflow.

Build current graph Recommended 
/ak:graphify .
Use when:
You need a first architecture map of an unfamiliar repository before planning.
Expected:
Builds graphify-out/graph.html, GRAPH_REPORT.md, graph.json, and cache artifacts, then uses the report to surface god nodes, surprising connections, and useful follow-up questions.
Report-focused analysis 
/ak:graphify src --report
Use when:
You mainly want the GRAPH_REPORT.md summary for architecture, dependency chains, and god-node review.
Expected:
Runs Graphify against src and centers the answer on GRAPH_REPORT.md findings: most-connected concepts, surprising relationships, and suggested questions for source inspection.
Prepare MCP queries 
/ak:graphify . --mcp
Use when:
You want Claude to query the generated graph across sessions through MCP graph tools.
Expected:
Builds graphify-out/graph.json and prepares MCP use with python -m graphify.serve so query_graph, get_node, get_neighbors, and shortest_path can inspect relationships.
Watch incremental rebuilds 
/ak:graphify . --watch
Use when:
You are changing files while exploring and want the graph refreshed as the project evolves.
Expected:
Runs watch mode and relies on the graphify-out/cache SHA256 hashes so rebuilds reprocess only changed files while keeping graph outputs current.

Handled Scope

  • Architecture discovery
  • Cross-file relationships
  • God-node analysis
  • MCP graph querying
  • Token-efficient context maps

Artifacts Produced

Graphify artifacts

Pattern

graphify-out/graph.html, GRAPH_REPORT.md, graph.json, cache/.

Use the report for orientation, the HTML for exploration, and graph.json/MCP for repeatable queries.

Related skills

Next