Skip to content
Back to skills

Zihin Haritasi

ASecurity

Turn any content into an interactive mind map (one HTML file). Use for "mind map", "map this", "zihin haritası", "haritala", "harita çiz", or a long answer the user wants at a glance.

  • 8 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added November 7, 2026
ai-agentspythongonodegit

Works with

  • claude code
  • terminal
  • cli

Security analysis

A100/100

Pro scans all 4 files and shows the line behind each finding

Scanned November 7, 2026

npx -y skills add leodavincs/harita --skill zihin-haritasi --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Zihin Haritasi?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Zihin Haritasi
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/leodavincs-zihin-haritasi/badge)](https://www.skillsdirectory.com/skills/leodavincs-zihin-haritasi)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: zihin-haritasi
description: Turn any content into an interactive mind map (one HTML file). Use for "mind map", "map this", "zihin haritası", "haritala", "harita çiz", or a long answer the user wants at a glance.
license: MIT
---

# Zihin haritası (mind map)

You turn content into an interactive mind map. You write the structure as JSON, the bundled script turns it into a single self-contained HTML page (no internet needed). The user sees the whole thing at a glance and opens details only when they want.

Paths below are relative to the folder that holds this SKILL.md.

## 1. Gather the content

- Default source: what is in this conversation (the topic, a long answer you gave, notes the user pasted, uploaded files).
- If the user asks to map things from earlier chats or "what you know about me", use whatever memory or past-chat search tools you have. If you have none, say so in one sentence and ask them to paste the material.
- Do not invent facts to fill branches. A short honest map beats a full fake one.

## 2. Pick a layout

| `duzen` | Use for |
| --- | --- |
| `harita` | default: report, research, plan, topic overview |
| `mantik` | step-by-step process, decision tree |
| `agac` | team, system parts, folder structure |
| `zaman` | week plan, roadmap, timeline (main branches in order) |
| `kilcik` | "why did this break": root = problem, main branches = causes |
| `matris` | comparison: main branch = row, children = cells |

## 3. Write the JSON

Read `references/bicim.md` once before your first map; it has the full format and a complete example. The keys are Turkish (`kok` = root, `alt` = children, `ad` = name, `not` = note); keep them exactly as written there.

Rules that make a good map:
- Write node text in the user's language.
- Node names are short: at most ~6 words. Everything longer goes into `not` (a string or a list of bullet strings), which opens when the node is clicked.
- 3-7 main branches. Depth 2-4. If a branch has more than ~8 children, group them.
- Root `not`: the one-sentence answer or conclusion, so the map is useful even without clicking.
- Use `isaret` markers sparingly: `karar` (a decision for the user), `kanit` (evidence), `fikir`, `uyari`, `soru`, `tamam`, `oncelik-1`…`oncelik-9`.
- Add `iliskiler` (arrows between distant nodes) only when the link is the point.
- Theme: leave `tema` out (default `beyaz`, clean white) unless the user asks for a look: `pastel` (soft, friendly), `kagit` (warm paper), `koyu` (dark), `gece` / `fosfor` / `amber` (terminal style). The viewer can also switch theme on the page; their choice is remembered.

## 4. Build the page

```
python3 scripts/kur.py <data.json> -o <output.html>
```

- The script checks the data first. If it prints `Veri hatalı:` followed by problems (duplicate id, unknown id, bad range, unknown layout), fix the JSON and run it again.
- Python 3.8+ standard library only; nothing to install.

## 5. Deliver

- **Claude app (claude.ai, desktop):** write the HTML into the outputs folder (`/mnt/user-data/outputs/` when it exists) so the user gets the file to open or download. If you can also show HTML as an artifact, you may show it there.
- **Local agents (Codex, Claude Code):** save next to the user's work (or the current folder) and add `--ac` to open it in the browser. Tell the user the file path.
- Never paste the generated HTML into the chat; it is large. Reply with one or two sentences: what the map shows and where it is.

## PDF

The page has a **PDF** button: page 1 is the map fitted to landscape A4, the following pages list every node with all its details. When the user asks for a PDF:
- Local agents: add `--pdf` to the build command; it writes `<output>.pdf` next to the HTML using Chrome, Chromium or Edge if installed. If none is found the script says so; then tell the user to open the page and press PDF.
- Claude app: try `--pdf` too; if the sandbox has no browser, deliver the HTML and tell the user to press the PDF button and choose "Save as PDF".

## 6. Changes

When the user asks for a change ("merge these two branches", "add a branch for X", "switch to timeline"), edit the same JSON and rebuild to the same output file. Keep `id`s stable so arrows and groups survive.

## When not to use a map

A short answer, a single fact, or text the user will copy and paste (a message, an email, code). Say so and answer normally.

Files in this skill

  • SKILL.md4.3 KB
  • assets/sablon.html60.2 KB
  • references/bicim.md3.2 KB
  • scripts/kur.py5.1 KB

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.