- name
- markdown-toc
- description
- Extract a Markdown heading tree with line numbers before editing or integrating notes. Use when you need to inspect a long Markdown file quickly and choose the best insertion point instead of editing blind.
# Markdown TOC
## Use this skill when
Use this skill before editing a long Markdown file, especially when you need to:
- inspect the heading structure quickly;
- choose the best section for new content;
- avoid creating redundant sections; or
- review nearby lines before patching a note.
## Quick start
Resolve `{baseDir}` to this skill’s actual directory, then run the bundled script. It reports ATX (`#`) headings, excluding fenced code; Setext headings and full Markdown parsing are outside its scope:
```bash
python3 "{baseDir}/scripts/extract_toc.py" /absolute/path/to/file.md
```
In this repository, common targets include:
```bash
python3 "{baseDir}/scripts/extract_toc.py" Notes/Gourmet.md
python3 "{baseDir}/scripts/extract_toc.py" "Notes/AI-Agent-Product&PE.md"
```
## Workflow
1. Run the TOC script on the candidate file.
2. Read the returned heading tree and line numbers.
3. Open the most likely section neighborhood with `sed -n '<start>,<end>p'` or `nl -ba`.
4. Insert into an existing section when possible.
5. Create a new section only when the file truly lacks a clean landing spot.
## Output format
The script prints a compact tree like this:
```text
TOC
--------------------------------------------------
# Title (line 1)
## Section (line 20)
### Subsection (line 42)
```
GitHub에서 보기