Imported from danielrosehill/Claude-Local-Web-Capture-Plugin (skills/scrape-authenticated/SKILL.md). Install upstream with npx skills add danielrosehill/Claude-Local-Web-Capture-Plugin --skill scrape-authenticated. Copyright stays with the author.

scrape-authenticated

Last-resort capture path. Drives the user's real Chrome, so requests carry the user's cookies, login state, and Israeli IP.

When to use

Only when one of the following is true:

• scrape-article reported that headless rungs 1–3 hit a paywall/login wall.
• sites.yaml marks the domain as strategy: authenticated (known paywalled site).
• User explicitly asks to capture from a logged-in session.

Otherwise reject and redirect to scrape-article. A GUI browser run is heavier, slower, and reveals the user's identity to the site.

Tooling

bb-browser (epiral/bb-browser) — CLI + MCP that drives the user's logged-in Chrome.

If bb-browser isn't installed, tell the user and recommend capture-setup rather than silently falling back.

Flow

1. Confirm the user is already logged in to the target site in Chrome. If not, ask them to sign in first; don't attempt auth flows from this skill.
2. Drive bb-browser to navigate to the URL, wait for the article container to render, then extract visible text.
3. Convert to markdown (same frontmatter/layout as scrape-article, with extractor: bb-browser and rung: 4).
4. Save to ~/local-web-capture/articles/YYYY/MM/ using the same filename convention.

Output

Same schema as scrape-article. Add to frontmatter:

extractor: bb-browser
rung: 4
authenticated: true

Do not

• Do not attempt to solve captchas or defeat protections. If the page challenges, stop and tell the user.
• Do not store cookies, credentials, or the page's Set-Cookie headers in the capture file.
• Do not run this for sites that worked on a lower rung — re-read sites.yaml first.