capture-webpage

name	capture-webpage
description	Use when archiving a URL as a self-contained offline HTML file with hash verification.
model	haiku

Capture Webpage

Archive a URL as a self-contained HTML file (via single-file), compute its hash, and report metadata.

When to use

• Capturing evidence web pages, online documents, or screenshots for chain-of-custody
• Archiving public web content before it changes or is removed
• Creating offline, tamper-evident copies of online material

Inputs to gather

• URL to capture (required)
• Destination directory (optional; defaults to current working directory)

Procedure

1. Invoke single-file <url> <dest_dir>/<sanitised-title>-<timestamp>.html where:
   • <sanitised-title> is the page title with spaces and special chars removed (or a sanitised URL slug)
   • <timestamp> is ISO-8601 or Unix epoch (e.g., 2026-04-30T22-18-00Z)
2. Wait for single-file to complete (may take 10–60 seconds depending on page complexity).
3. After success, compute sha256sum <output.html> and capture the hash.
4. Report to user: output file path, SHA-256 hash, file size in bytes, and final URL (after redirects, as captured by single-file's --browser-args).
5. Suggest piping the hash + path through /legal-investigative log-evidence for custody registration.

Output / side effects

• <dest_dir>/<sanitised-title>-<timestamp>.html — self-contained webpage archive.
• User sees: output path, SHA-256 hash, byte size.
• No custody log entry written (suggest user chain through legal-investigative's /log-evidence for that).

Safety / constraints

• single-file requires a browser (Chrome/Chromium) to be available; ensure it's installed.
• URLs with invalid characters or redirects may produce unexpected output filenames; show user the final path.
• Large, media-heavy pages may produce very large HTML files; warn if >100 MB.