Skill Graphify

该 Skill 通过跨平台 Python 包装脚本调用 graphify CLI，对指定文件夹执行文件识别、AST 抽取、建图、聚类与导出，产出交互式 HTML、可查询 JSON 图谱和白话审计报告，支持图谱查询、增量缓存更新与忽略规则，语义抽取依赖可用的 LLM 子代理。

flobo3

热度 64.72

最近更新 2026-05-11 17:41:50

分类：知识依赖：代码执行 / 读写文件 / 内部调用大模型应用场景：研究 / 报告撰写 / 摘要总结 / 笔记整理 ...

安装技能

复制下面的话术发给你的 Agent，它会通过 secure-skills 检索并安装该技能。

执行命令安装此 skill： npx secure-skills-cli add skill-graphify

Metadata

name: skill-graphify
description: Turn any folder of code, docs, papers, or images into a queryable knowledge graph. Cross-platform wrapper for graphify CLI.

Skill Graphify

Turn any folder of files into a navigable knowledge graph with community detection, honest audit trail, and three outputs: interactive HTML, queryable JSON, and a plain-language report.

When to use

User wants to understand a codebase they're new to
User asks "how does X connect to Y" across many files
User has a folder of papers/notes/screenshots and wants structure
User wants a visual map of their project's architecture

Usage

Step 1 — Ensure graphify is installed

python graphify_wrapper.py ensure-installed

Or manually: pip install graphifyy

Step 2 — Build knowledge graph

python graphify_wrapper.py build /path/to/project

This runs the full pipeline: detect files → AST extraction → build graph → cluster → export.

Output goes to <project>/graphify-out/:

graph.html — interactive visualization (open in browser)
GRAPH_REPORT.md — plain-language audit report
graph.json — queryable knowledge graph
cache/ — SHA256 cache for incremental updates

Step 3 — Read the report

python graphify_wrapper.py report

Or read graphify-out/GRAPH_REPORT.md directly. Present the key findings to the user: god nodes (highly connected), surprising connections, community structure.

Step 4 — Query the graph (optional)

python graphify_wrapper.py query "how does authentication work"

Or use the CLI directly for more options:

graphify query "show the auth flow" --graph graphify-out/graph.json
graphify query "what connects X to Y?" --graph graphify-out/graph.json --dfs
graphify query "explain dependency injection" --budget 1500 --graph graphify-out/graph.json

Send results to user

After building, send graphify-out/graph.html to the user so they can explore the interactive graph. Summarize GRAPH_REPORT.md in your response.

CLI reference (graphify)

If graphify CLI is available, you can use these directly:

Command	Description
`graphify query "..." --graph <path>`	BFS traversal of the graph
`graphify query "..." --dfs --graph <path>`	DFS — trace a specific path
`graphify query "..." --budget N --graph <path>`	Cap output at N tokens
`graphify path "Node1" "Node2" --graph <path>`	Shortest path between concepts
`graphify explain "NodeName" --graph <path>`	Plain-language explanation of a node

Supported file types

Code: 20 languages via tree-sitter (Python, JS, TS, Go, Rust, Java, C, C++, Ruby, C#, Kotlin, Scala, PHP, Swift, Lua, Zig, PowerShell, Elixir, Objective-C, Julia)
Docs: Markdown, text, reStructuredText
Papers: PDF
Images: Screenshots, diagrams, whiteboard photos (requires vision-capable LLM)

Notes

The wrapper script (graphify_wrapper.py) handles cross-platform compatibility (Windows CMD, Linux, macOS)
graphify's AST extraction is deterministic and requires no LLM — it's free and fast
Semantic extraction (docs, images) uses LLM subagents if available, otherwise is skipped
Every edge is tagged EXTRACTED, INFERRED, or AMBIGUOUS — you always know what was found vs guessed
Incremental updates: re-running on the same folder only processes changed files (cache-based)
Add a .graphifyignore file (same syntax as .gitignore) to exclude directories

Dependencies

Python 3.10+
graphifyy (pip) — automatically installed by wrapper