Reverse-engineer a website's internal APIs, encrypted endpoints, WebSocket streams, and obfuscated JavaScript. Activates when the target data isn't in the HTML, when the site uses encrypted CloudFront/CDN payloads, when real-time streaming data is needed, or when the scrape skill's Phase 0-1 finds API calls that are encrypted, signed, or behind a custom protocol. Escalation ladder from simple network capture to protobuf schema reconstruction.
Scrape a website. Activates when user provides a URL to scrape, asks to extract data from a site, hits 403/blocking errors, or needs anti-bot evasion.
Give AI agents their own email inboxes using the AgentMail API. Use when building email agents, sending/receiving emails programmatically, managing inboxes, handling attachments, organizing with labels, creating drafts for human approval, or setting up real-time notifications via webhooks/websockets. Supports multi-tenant isolation with pods.
Use Camoufox — a Firefox fork with C++-level fingerprint spoofing — for browser scraping when the user's real Chrome profile is not available. Activate on VMs, headless CI servers, remote machines without a GUI, or when torch needs multiple concurrent personas with rotated fingerprints. Detects via TORCH_CAMOUFOX_ENDPOINT env var. Drives from Node via playwright-core connecting to a Camoufox-hosted Playwright server.
Proven scraping playbook for amazon.com search result pages (/s?k=...). CloudFront + CAPTCHA wall blocks bare curl with HTTP 503, but a real Chrome session via the real Chrome debug port walks right through — no stealth, no proxy, no captcha. HTML is server-rendered; cheerio parses 22 results per page with stable `[data-component-type="s-search-result"]` blocks. Activate for any amazon.com /s search URL.
Proven scraping playbook for booking.com searchresults.html pages. CloudFront-fronted JS challenge blocks bare curl (HTTP 202 with a script-only interstitial), but a real Chrome session via the real Chrome debug port walks through on first navigation — no captcha, no proxy, no login. Listings are rendered client-side into [data-testid="property-card"] cards; pagination uses a "Load more results" button after the first 25. Prices only populate when the URL carries checkin/checkout/group_adults params. Activate for any booking.com /searchresults.html scrape.
Proven scraping playbook for costco.com category listings (e.g. /laptops.html, /computers.html). Akamai Bot Manager hard-blocks bare curl with HTTP 403 on the HTML pages, but Costco's internal catalog API at gdx-api.costco.com/catalog/search/api/v1/search is a public JSON POST endpoint with no auth, no cookies, and no anti-bot. Skip the browser entirely. Activate for any costco.com category listing scrape.
Proven scraping playbook for digikey.com /en/products/category/ listings. Cloudflare challenge (cf-mitigated: challenge) blocks bare curl with HTTP 403, but a puppeteer connection to the user's real Chrome via the real Chrome debug port walks straight through with zero stealth, no captcha, no proxy, no auth. Listings are server-rendered into a single `<table>` whose rows contain `a[href*="/en/products/detail/"]`; every parametric column (Core Processor, Speed, RAM, package, etc.) is just a `<th>`/`<td>` pair you can zip. Activate for any digikey.com /en/products/category/ scrape.