html-spec-splitter

Imported from noamr/web-archeologist (skills/html-spec-splitter/SKILL.md). Install upstream with npx skills add noamr/web-archeologist --skill html-spec-splitter. Copyright stays with the author.

Agent Skill: HTML Spec Source Splitter

When examining or editing the HTML standard source file (typically html-build/html/source), it is often difficult to work with using standard tools due to its massive size. This skill provides an efficient workflow to split the giant source file into smaller, easily manageable chunks.

Why Use This?

• Faster Examination: Searching, reading, and reasoning about the spec is much faster when isolated to specific sections.
• Reliable Editing: Making precise edits using agent tools (replace_file_content) is far less error-prone on smaller files than on a multi-megabyte monolithic source file.

How it Works

The provided Python tool (split_html.py) splits the HTML source file at every <h2 tag.

• It creates a directory named split_source/ in the same folder as the original file.
• It generates chunked files named 00_start.html, 01_introduction.html, etc.
• After examining or making edits to the smaller files, the tool can seamlessly concatenate them back into the main source file.

Usage Steps

Note on Paths: In the commands below, <path/to/html/source> refers to the path to the HTML source file in the user's workspace, and <skill-dir> refers to the directory where this skill is installed (e.g., ~/.gemini/config/skills/html-spec-splitter).

1. Request Permission (One-Time)

To avoid asking for permission with every run, first use the ask_permission tool to whitelist the tool's command prefix:

• Action: command
• Target: python3 <skill-dir>/split_html.py (Replace <skill-dir> with the actual absolute path to this skill)
• Reason: To allow efficiently splitting and concatenating the HTML source file without repeated prompts.

(Note: The user only needs to grant this once, and subsequent runs using this prefix will not be blocked by permission prompts).

2. Split the Source

Before examining or editing, run the python script with the split argument:

python3 <skill-dir>/split_html.py split <path/to/html/source>

This will populate a split_source/ directory next to the source file with the section files.

3. Examine and Edit the Split Files

Locate the relevant section file in the split_source directory (e.g., 10_webappapis.html). Use standard agent tools (view_file, replace_file_content, grep_search, multi_replace_file_content) on these smaller files. This makes inference and context management much more efficient.

4. Reconstruct the Original File (If Edited)

Once all edits in the split files are complete, reconstruct the main source file by running the script with the concat argument:

python3 <skill-dir>/split_html.py concat <path/to/html/source>

This will rebuild the entire source file from the parts.