Skip to content
Skills Library
Programming · Analysis

graphify

/graphify

Use when exploring a big or unfamiliar codebase, or asked how its parts connect or depend on each other: builds a queryable code knowledge graph with the graphify CLI.

Licence
MIT
Last repository change
13.11.2026

Assessment and security

100/ 100Class: good
Format25 / 25
Frontmatter with name and description, valid name, description up to 1,024 characters, name matching the folder.
Discoverability25 / 25
Description explains when to use the skill, has a reasonable length and uses third person.
Content35 / 35
At least 150 words and at most 500 lines, headings, step lists, examples or code.
Hygiene15 / 15
No placeholders such as TODO, no Windows paths, with a licence.
Riskno signals detected in the scan
Files in this package
1
Scripts
0
HTML comments
0
URLs in SKILL.md
0
Deprecation

No signals detected. Repository changed on 13.11.2026, without references to retired models or APIs.

Licence

MIT, open: use and modify while retaining author attribution.

What it does and when to use it

Use when exploring a big or unfamiliar codebase, or asked how its parts connect or depend on each other: builds a queryable code knowledge graph with the graphify CLI.

Original author description

Use when exploring a big or unfamiliar codebase, or asked how its parts connect or depend on each other: builds a queryable code knowledge graph with the graphify CLI.

How it works

  1. When to use
  2. First, check that it is there
  3. Build the graph (code only, no key, no tokens)
  4. Outputs (graphify-out/)
  5. Ask the graph
  6. Order of work
Original SKILL.md headings · 47 lines, 570 words
SKILL.md preview
---
name: graphify
description: "Use when exploring a big or unfamiliar codebase, or asked how its parts connect or depend on each other: builds a queryable code knowledge graph with the graphify CLI."
version: 1.0.0
category: dev
author: Graphify Labs (adapted for Macron)
license: MIT
---

# graphify: a codebase as a knowledge graph

`graphify` parses a folder of code into a graph (functions, classes, imports, calls), groups it into communities of related code, and answers questions from the graph instead of from dozens of files. In the machine it runs on the code alone, with no model and no API key, so a question costs almost nothing.

## When to use
- A new or big codebase, or several services, where `project_tree` and `code_outline` are not enough.
- Questions about structure, dependencies or relationships across many files: "what calls X", "how does A reach B", "what are the main parts".
- Not for one file or one function: read it, or use `code_outline` and `search_code`.

## First, check that it is there
1. `bash_execute`: `command -v graphify`. If it prints nothing, install it into the user's environment with `install_package` (manager `pip`, package **`graphifyy`** with two y's). It adds the `graphify` command for `bash_execute`.
2. If it cannot be installed (no network, an error), say so in one line and fall back to `project_tree`, `code_outline` and `search_code`. Never pretend a graph was built.
3. Never put an API key here for it, and never run `graphify install`, `graphify global` or the MCP server: none is needed.

## Build the graph (code only, no key, no tokens)
```bash
graphify <folder> --code-only                          # AST of the code → graphify-out/graph.json
GRAPHIFY_BACKEND=none graphify cluster-only <folder>   # → GRAPH_REPORT.md and graph.html, communities named "Community N"
```
- Plain `graphify .` also reads docs, papers and images with a model and stops asking for an API key: always pass `--code-only`. `GRAPHIFY_BACKEND=none` keeps the second step from calling any model.
- A GitHub repository: bring it into the workspace first with `import_repo`, then build on its folder (the machine command has no network).
- A big repository takes a while: run it as a background job and look at its log (`bash_execute` with background, `shell_jobs`).
- After the code changed: `graphify update <folder>` re-extracts the code (no model); rebuild the report with `cluster-only` if you need it.

## Outputs (`graphify-out/`)
1. `GRAPH_REPORT.md`: communities, hubs, the most connected nodes, suggested queries. **Read it first.**
2. `graph.json`: the whole graph; everything below queries it.
3. `graph.html`: an interactive picture of the graph; send it with `send_file` when the user wants to see it.

## Ask the graph
```bash
graphify query "how are tools registered and executed"   # broad context (BFS)
graphify query "how are tools registered" --dfs          # follow one chain deeply
graphify query "…" --budget 5000                         # more nodes when the answer says it was truncated
graphify path "Registry" "Machine"                       # the shortest path between two nodes
graphify explain "Registry"                              # one node and its connections
```
- `query`, `path` and `explain` exit 3 when a name is ambiguous: use the exact label from the report or from a first `query`.
- Names are the code's own (functions, types, packages); quote them exactly.

## Order of work
1. If `graphify-out/graph.json` is missing or the code changed, build it as above.
2. Read `GRAPH_REPORT.md` for the architecture and the communities.
3. Answer with `query`, `path` and `explain`; open the real files (`read_file`, `search_code`) only for what you will quote or change.
4. Say which parts of the answer come from the graph and which you checked in the source. A graph shows structure, not behavior: the edges are extracted or inferred, so confirm what matters.

Versions

RepositoryScoreLicence
This page · Metixpro/Macron-Agent100MIT
vanducng/skills100MIT
TerminalSkills/skills100Apache-2.0
2799662352/ai-image-master100MIT
adeiltonpessini/kit-memoria-ia100Apache-2.0
vinnie357/claude-skills100MIT
v1truv1us/ai-eng-system100MIT
sammcj/agentic-coding95Apache-2.0
taianh20102011/skill-for-ai95MIT

Database snapshot: 13.11.2026 · package version 5fa5fc55c9af (SHA-256 digest of the entire package)

Downloads in the last 30 days: 0 · ranking

Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.

Similar skills

  • 100graphify-codebase
    Graphify(コードベースをナレッジグラフ化するCLI)を導入・更新・照会し、SDDの設計・レビュー・実装・トラブルシューティングで影響範囲や依存関係の調査に使う。既存コードの構造把握、変更の波及範囲の特定、2要素間の経路追跡が必要な場合に使用する。Graphifyは任意の補助ツールであり、未導入でも既存の調査手順にフォールバックする。Do NOT use for 要求の整合性検証(requirement-managementのreqctl.pyを使用すること)。
  • 100graphify-usage
    This skill should be used when querying the graphify knowledge graph for structural codebase information, choosing between graph tools and grep, or interpreting graph query results.
  • 90project-knowledge-graph
    基于 graphify(Graphify-Labs/graphify,多模态知识图谱生成器)做项目/语料的知识图谱 构建:当用户接手陌生大型代码库要先摸清结构枢纽、要把代码+文档+论文+截图等异构语料 织成统一概念网络、或要给后续 agent 会话留下可增量更新的 wiki/图谱导航层时触发。 边界:不做动态对话记忆与用户画像、不做经验 SOP 沉淀;小语料(≤6 文件)收益有限需 如实告知。
  • 90explain-code
    Use when the owner wants a code project mapped as a knowledge graph or wants code explained to a beginner, using Understand-Anything or graphify.
  • 100graph
    Use when the user wants to build, refresh, or visualize the repo's knowledge graph — code structure plus the committed wiki/ — via graphify. Produces an interactive graph + a GRAPH_REPORT.md of God Nodes, surprising connections, and suggested questions. Trigger on "/graph",…
  • 86graphify-windows
    any input (code, docs, papers, images) → knowledge graph → clustered communities → HTML + JSON + audit report. Use when user asks any question about a codebase, project content, architecture, or file relationships — especially if graphify-out/ exists. Provides persistent graph…

Feedback and reports

Your message goes to Bartosz Sokoliński. An email address is needed only for a reply. See the privacy policy.