What to do next on this site

You are working on ONE specific website: stumblicious.com (live at https://stumblicious.com/), a Web Curation site in the WholeTech network, served from /var/www/stumblicious.com/ on the droplet (ssh root@143.198.182.180). Every instruction below is for stumblicious.com and no other site — make all changes on stumblicious.com. Act like a senior editor and web engineer who knows the Web Curation niche cold.

GOAL (on stumblicious.com)
Make the site mythos-attack-proof — harden against automated & AI-agent abuse — on the site stumblicious.com.

WHY IT MATTERS
Assume a determined automated, AI-driven attacker. Enforce HTTPS, add a sensible Content-Security-Policy plus the standard security headers, eliminate open redirects and reflected user input, validate and rate-limit any form, and ensure every agent/API endpoint and machine file (AGENTS.md, llms.txt, *.json) returns ONLY public data — never internal or personal information.

HOW TO APPROACH IT
1. Confirm nothing sensitive is publicly reachable: as a logged-out visitor, /.git/, /.env, backups (.bak/.zip/.sql/.tgz), config files, .htpasswd, and any private or admin path must return 403 or 404 — never 200.
2. Scan the served HTML, JS, and JSON for leaked secrets, API keys, tokens, internal emails, or personal data, and remove anything that should not be public.
3. Enforce HTTPS and confirm the standard security headers are present (HSTS, X-Frame-Options, X-Content-Type-Options, Referrer-Policy, and a sensible Content-Security-Policy).
4. Harden anything interactive against a determined automated or AI-driven attacker: validate and rate-limit forms, eliminate open redirects and reflected input, and make sure every agent/API endpoint and machine file (AGENTS.md, llms.txt, *.json) returns ONLY public data.

CONSTRAINTS
- Match the site's existing design, fonts, layout and voice — never a generic AI-template look.
- Keep the HTML valid; do not break nav, internal links, the promo bar, analytics, or ad units.
- Use only real, current, accurately-attributed information — no fabricated facts, numbers, or sources.
- Reuse the site's existing Google Analytics and AdSense.

WHAT "DONE" LOOKS LIKE
- A logged-out request to every sensitive path returns 403/404 — verified, not assumed.
- No secret, key, token, internal email, or personal datum appears anywhere in the served HTML, JS, or JSON.
- HTTPS is forced and security headers are confirmed live; the Agents-First surface exposes only public information — the site is mythos-attack-proof.
- Deployed to the droplet and verified on the live URL (screenshot from a Windows machine — never run a browser on the droplet), with a short summary of exactly what changed.

Think through the best approach first, then complete the whole task autonomously before stopping.

You are working on ONE specific website: stumblicious.com (live at https://stumblicious.com/), a Web Curation site in the WholeTech network, served from /var/www/stumblicious.com/ on the droplet (ssh root@143.198.182.180). Every instruction below is for stumblicious.com and no other site — make all changes on stumblicious.com. Act like a senior editor and web engineer who knows the Web Curation niche cold.

GOAL (on stumblicious.com)
Lock down sensitive data — nothing private is publicly reachable — on the site stumblicious.com.

WHY IT MATTERS
Make sure a logged-out visitor or bot cannot reach anything sensitive on this site: /.git/, /.env, backups (.bak/.zip/.sql/.tgz), config files, .htpasswd, and any private or admin path must return 403 or 404, never 200. Then scan the served HTML, JS, and JSON for leaked keys, tokens, internal emails, or personal data and remove them.

HOW TO APPROACH IT
1. Confirm nothing sensitive is publicly reachable: as a logged-out visitor, /.git/, /.env, backups (.bak/.zip/.sql/.tgz), config files, .htpasswd, and any private or admin path must return 403 or 404 — never 200.
2. Scan the served HTML, JS, and JSON for leaked secrets, API keys, tokens, internal emails, or personal data, and remove anything that should not be public.
3. Enforce HTTPS and confirm the standard security headers are present (HSTS, X-Frame-Options, X-Content-Type-Options, Referrer-Policy, and a sensible Content-Security-Policy).
4. Harden anything interactive against a determined automated or AI-driven attacker: validate and rate-limit forms, eliminate open redirects and reflected input, and make sure every agent/API endpoint and machine file (AGENTS.md, llms.txt, *.json) returns ONLY public data.

CONSTRAINTS
- Match the site's existing design, fonts, layout and voice — never a generic AI-template look.
- Keep the HTML valid; do not break nav, internal links, the promo bar, analytics, or ad units.
- Use only real, current, accurately-attributed information — no fabricated facts, numbers, or sources.
- Reuse the site's existing Google Analytics and AdSense.

WHAT "DONE" LOOKS LIKE
- A logged-out request to every sensitive path returns 403/404 — verified, not assumed.
- No secret, key, token, internal email, or personal datum appears anywhere in the served HTML, JS, or JSON.
- HTTPS is forced and security headers are confirmed live; the Agents-First surface exposes only public information — the site is mythos-attack-proof.
- Deployed to the droplet and verified on the live URL (screenshot from a Windows machine — never run a browser on the droplet), with a short summary of exactly what changed.

Think through the best approach first, then complete the whole task autonomously before stopping.

You are working on ONE specific website: stumblicious.com (live at https://stumblicious.com/), a Web Curation site in the WholeTech network, served from /var/www/stumblicious.com/ on the droplet (ssh root@143.198.182.180). Every instruction below is for stumblicious.com and no other site — make all changes on stumblicious.com. Act like a senior editor and web engineer who knows the Web Curation niche cold.

GOAL (on stumblicious.com)
Add an interactive tool and an email capture for Web Curation — on the site stumblicious.com.

WHY IT MATTERS
Build one genuinely useful interactive element for Web Curation (a calculator, checklist, finder, or quiz) plus a tasteful email capture. Tools earn repeat visits, backlinks, and a first-party audience the site actually owns.

HOW TO APPROACH IT
1. Study the strongest existing pages on this topic anywhere and decide what would make THIS the single most useful page in Web Curation — then build exactly that.
2. Write substantial, original content with a clear structure: a strong thesis up top, well-organized sections, concrete real examples and current numbers, and a decisive takeaway.
3. Ground every claim in current, authoritative sources and link out to them inline; never invent facts or statistics.
4. Wire it into the site: link from the homepage and nav, cross-link /sources/ and related pages, and add Article + FAQ schema.

CONSTRAINTS
- Match the site's existing design, fonts, layout and voice — never a generic AI-template look.
- Keep the HTML valid; do not break nav, internal links, the promo bar, analytics, or ad units.
- Use only real, current, accurately-attributed information — no fabricated facts, numbers, or sources.
- Reuse the site's existing Google Analytics and AdSense.

WHAT "DONE" LOOKS LIKE
- The page is genuinely the strongest Web Curation resource on the site — deep, accurate, and a pleasure to read, not a thin template.
- Every factual claim is current and traceable to a real, linked source.
- It's discoverable: linked from the homepage, in the sitemap, and submitted via IndexNow.
- Deployed to the droplet and verified on the live URL (screenshot from a Windows machine — never run a browser on the droplet), with a short summary of exactly what changed.

Think through the best approach first, then complete the whole task autonomously before stopping.

You are working on ONE specific website: stumblicious.com (live at https://stumblicious.com/), a Web Curation site in the WholeTech network, served from /var/www/stumblicious.com/ on the droplet (ssh root@143.198.182.180). Every instruction below is for stumblicious.com and no other site — make all changes on stumblicious.com. Act like a senior editor and web engineer who knows the Web Curation niche cold.

GOAL (on stumblicious.com)
Verify HTTPS, security headers & a safe Agents-First surface — on the site stumblicious.com.

WHY IT MATTERS
Confirm the certificate is valid and HTTPS is forced, the standard security headers are present (HSTS, X-Frame-Options, X-Content-Type-Options, Referrer-Policy, CSP), and the Agents-First endpoints expose only public, non-sensitive information — trustworthy to humans and safe for the agents reading it.

HOW TO APPROACH IT
1. Confirm nothing sensitive is publicly reachable: as a logged-out visitor, /.git/, /.env, backups (.bak/.zip/.sql/.tgz), config files, .htpasswd, and any private or admin path must return 403 or 404 — never 200.
2. Scan the served HTML, JS, and JSON for leaked secrets, API keys, tokens, internal emails, or personal data, and remove anything that should not be public.
3. Enforce HTTPS and confirm the standard security headers are present (HSTS, X-Frame-Options, X-Content-Type-Options, Referrer-Policy, and a sensible Content-Security-Policy).
4. Harden anything interactive against a determined automated or AI-driven attacker: validate and rate-limit forms, eliminate open redirects and reflected input, and make sure every agent/API endpoint and machine file (AGENTS.md, llms.txt, *.json) returns ONLY public data.

CONSTRAINTS
- Match the site's existing design, fonts, layout and voice — never a generic AI-template look.
- Keep the HTML valid; do not break nav, internal links, the promo bar, analytics, or ad units.
- Use only real, current, accurately-attributed information — no fabricated facts, numbers, or sources.
- Reuse the site's existing Google Analytics and AdSense.

WHAT "DONE" LOOKS LIKE
- A logged-out request to every sensitive path returns 403/404 — verified, not assumed.
- No secret, key, token, internal email, or personal datum appears anywhere in the served HTML, JS, or JSON.
- HTTPS is forced and security headers are confirmed live; the Agents-First surface exposes only public information — the site is mythos-attack-proof.
- Deployed to the droplet and verified on the live URL (screenshot from a Windows machine — never run a browser on the droplet), with a short summary of exactly what changed.

Think through the best approach first, then complete the whole task autonomously before stopping.

You are working on ONE specific website: stumblicious.com (live at https://stumblicious.com/), a Web Curation site in the WholeTech network, served from /var/www/stumblicious.com/ on the droplet (ssh root@143.198.182.180). Every instruction below is for stumblicious.com and no other site — make all changes on stumblicious.com. Act like a senior editor and web engineer who knows the Web Curation niche cold.

GOAL (on stumblicious.com)
Build the definitive Web Curation cornerstone guide — on the site stumblicious.com.

WHY IT MATTERS
Create the single most useful, authoritative page on Web Curation anywhere — the resource the whole site points to. Deep, current, expertly structured, and link-worthy. This anchors the site's authority and is the page most likely to earn rankings and citations.

HOW TO APPROACH IT
1. Study the strongest existing pages on this topic anywhere and decide what would make THIS the single most useful page in Web Curation — then build exactly that.
2. Write substantial, original content with a clear structure: a strong thesis up top, well-organized sections, concrete real examples and current numbers, and a decisive takeaway.
3. Ground every claim in current, authoritative sources and link out to them inline; never invent facts or statistics.
4. Wire it into the site: link from the homepage and nav, cross-link /sources/ and related pages, and add Article + FAQ schema.

CONSTRAINTS
- Match the site's existing design, fonts, layout and voice — never a generic AI-template look.
- Keep the HTML valid; do not break nav, internal links, the promo bar, analytics, or ad units.
- Use only real, current, accurately-attributed information — no fabricated facts, numbers, or sources.
- Reuse the site's existing Google Analytics and AdSense.

WHAT "DONE" LOOKS LIKE
- The page is genuinely the strongest Web Curation resource on the site — deep, accurate, and a pleasure to read, not a thin template.
- Every factual claim is current and traceable to a real, linked source.
- It's discoverable: linked from the homepage, in the sitemap, and submitted via IndexNow.
- Deployed to the droplet and verified on the live URL (screenshot from a Windows machine — never run a browser on the droplet), with a short summary of exactly what changed.

Think through the best approach first, then complete the whole task autonomously before stopping.

You are working on ONE specific website: stumblicious.com (live at https://stumblicious.com/), a Web Curation site in the WholeTech network, served from /var/www/stumblicious.com/ on the droplet (ssh root@143.198.182.180). Every instruction below is for stumblicious.com and no other site — make all changes on stumblicious.com. Act like a senior editor and web engineer who knows the Web Curation niche cold.

GOAL (on stumblicious.com)
Make the homepage genuinely distinctive — kill any generic-template feel — on the site stumblicious.com.

WHY IT MATTERS
Give the homepage a clear point of view, a memorable hero, and real utility above the fold. A distinctive, intentional design lifts trust, time-on-page, and the Agents-First score out of generic-template territory.

HOW TO APPROACH IT
1. Study the strongest existing pages on this topic anywhere and decide what would make THIS the single most useful page in Web Curation — then build exactly that.
2. Write substantial, original content with a clear structure: a strong thesis up top, well-organized sections, concrete real examples and current numbers, and a decisive takeaway.
3. Ground every claim in current, authoritative sources and link out to them inline; never invent facts or statistics.
4. Wire it into the site: link from the homepage and nav, cross-link /sources/ and related pages, and add Article + FAQ schema.

CONSTRAINTS
- Match the site's existing design, fonts, layout and voice — never a generic AI-template look.
- Keep the HTML valid; do not break nav, internal links, the promo bar, analytics, or ad units.
- Use only real, current, accurately-attributed information — no fabricated facts, numbers, or sources.
- Reuse the site's existing Google Analytics and AdSense.

WHAT "DONE" LOOKS LIKE
- The page is genuinely the strongest Web Curation resource on the site — deep, accurate, and a pleasure to read, not a thin template.
- Every factual claim is current and traceable to a real, linked source.
- It's discoverable: linked from the homepage, in the sitemap, and submitted via IndexNow.
- Deployed to the droplet and verified on the live URL (screenshot from a Windows machine — never run a browser on the droplet), with a short summary of exactly what changed.

Think through the best approach first, then complete the whole task autonomously before stopping.

You are working on ONE specific website: stumblicious.com (live at https://stumblicious.com/), a Web Curation site in the WholeTech network, served from /var/www/stumblicious.com/ on the droplet (ssh root@143.198.182.180). Every instruction below is for stumblicious.com and no other site — make all changes on stumblicious.com. Act like a senior editor and web engineer who knows the Web Curation niche cold.

GOAL (on stumblicious.com)
Daily five-link stumble with one-line takes — on the site stumblicious.com.

WHY IT MATTERS
Publish a dated daily page of five hand-picked odd-and-wonderful web finds, each with a single sharp sentence of commentary — never mirrored copy. Daily cadence builds a check-in habit, dated URLs accumulate long-tail coverage, and an RSS feed puts stumblicious into the small-web reader ecosystem where curation sites earn their links. (Fable enhanced action — also on the global worksheet.)

HOW TO APPROACH IT
1. Plan the cleanest way to accomplish this, then execute it well.
2. Use real, current information and match the existing quality of the site.

CONSTRAINTS
- Match the site's existing design, fonts, layout and voice — never a generic AI-template look.
- Keep the HTML valid; do not break nav, internal links, the promo bar, analytics, or ad units.
- Use only real, current, accurately-attributed information — no fabricated facts, numbers, or sources.
- Reuse the site's existing Google Analytics and AdSense.

WHAT "DONE" LOOKS LIKE
- The task is fully done, accurate, and live.
- Deployed to the droplet and verified on the live URL (screenshot from a Windows machine — never run a browser on the droplet), with a short summary of exactly what changed.

Think through the best approach first, then complete the whole task autonomously before stopping.

You are working on ONE specific website: stumblicious.com (live at https://stumblicious.com/), a Web Curation site in the WholeTech network, served from /var/www/stumblicious.com/ on the droplet (ssh root@143.198.182.180). Every instruction below is for stumblicious.com and no other site — make all changes on stumblicious.com. Act like a senior editor and web engineer who knows the Web Curation niche cold.

GOAL (on stumblicious.com)
Publish: The question your Web Curation audience is searching most right now — on the site stumblicious.com.

WHY IT MATTERS
Trending angle in Web Curation. A fresh, data-backed page on this draws search and gives the site a current spine. Link it from the homepage and /sources/.

HOW TO APPROACH IT
1. Research the angle right now using at least 3 authoritative, current sources; capture the real, specific data points.
2. Write a focused page: a clear thesis, the evidence, what it concretely means for a Web Curation reader, and a clear next step.
3. Cite and link each source inline and date the piece so its freshness is obvious.
4. Link it from the homepage and /sources/, and submit it via IndexNow.

CONSTRAINTS
- Match the site's existing design, fonts, layout and voice — never a generic AI-template look.
- Keep the HTML valid; do not break nav, internal links, the promo bar, analytics, or ad units.
- Use only real, current, accurately-attributed information — no fabricated facts, numbers, or sources.
- Reuse the site's existing Google Analytics and AdSense.

WHAT "DONE" LOOKS LIKE
- A reader learns something current and genuinely useful they could not get from a thin AI page.
- At least 3 real sources are cited and linked; no fabricated numbers.
- The page is internally linked and submitted for fast indexing.
- Deployed to the droplet and verified on the live URL (screenshot from a Windows machine — never run a browser on the droplet), with a short summary of exactly what changed.

Think through the best approach first, then complete the whole task autonomously before stopping.

You are working on ONE specific website: stumblicious.com (live at https://stumblicious.com/), a Web Curation site in the WholeTech network, served from /var/www/stumblicious.com/ on the droplet (ssh root@143.198.182.180). Every instruction below is for stumblicious.com and no other site — make all changes on stumblicious.com. Act like a senior editor and web engineer who knows the Web Curation niche cold.

GOAL (on stumblicious.com)
Publish: A current, data-backed explainer in Web Curation — on the site stumblicious.com.

WHY IT MATTERS
Trending angle in Web Curation. A fresh, data-backed page on this draws search and gives the site a current spine. Link it from the homepage and /sources/.

HOW TO APPROACH IT
1. Research the angle right now using at least 3 authoritative, current sources; capture the real, specific data points.
2. Write a focused page: a clear thesis, the evidence, what it concretely means for a Web Curation reader, and a clear next step.
3. Cite and link each source inline and date the piece so its freshness is obvious.
4. Link it from the homepage and /sources/, and submit it via IndexNow.

CONSTRAINTS
- Match the site's existing design, fonts, layout and voice — never a generic AI-template look.
- Keep the HTML valid; do not break nav, internal links, the promo bar, analytics, or ad units.
- Use only real, current, accurately-attributed information — no fabricated facts, numbers, or sources.
- Reuse the site's existing Google Analytics and AdSense.

WHAT "DONE" LOOKS LIKE
- A reader learns something current and genuinely useful they could not get from a thin AI page.
- At least 3 real sources are cited and linked; no fabricated numbers.
- The page is internally linked and submitted for fast indexing.
- Deployed to the droplet and verified on the live URL (screenshot from a Windows machine — never run a browser on the droplet), with a short summary of exactly what changed.

Think through the best approach first, then complete the whole task autonomously before stopping.

You are working on ONE specific website: stumblicious.com (live at https://stumblicious.com/), a Web Curation site in the WholeTech network, served from /var/www/stumblicious.com/ on the droplet (ssh root@143.198.182.180). Every instruction below is for stumblicious.com and no other site — make all changes on stumblicious.com. Act like a senior editor and web engineer who knows the Web Curation niche cold.

GOAL (on stumblicious.com)
Mood-based small-web directory with chips — on the site stumblicious.com.

WHY IT MATTERS
Build a browsable directory of vetted independent sites tagged by mood — 'rabbit hole', 'soothing', 'weird tools', 'time capsule' — instead of dry categories. Mood navigation is the StumbleUpon feeling nobody has recreated, and small-web owners reliably link back to directories that feature them thoughtfully. (Fable enhanced action — also on the global worksheet.)

HOW TO APPROACH IT
1. Plan the cleanest way to accomplish this, then execute it well.
2. Use real, current information and match the existing quality of the site.

CONSTRAINTS
- Match the site's existing design, fonts, layout and voice — never a generic AI-template look.
- Keep the HTML valid; do not break nav, internal links, the promo bar, analytics, or ad units.
- Use only real, current, accurately-attributed information — no fabricated facts, numbers, or sources.
- Reuse the site's existing Google Analytics and AdSense.

WHAT "DONE" LOOKS LIKE
- The task is fully done, accurate, and live.
- Deployed to the droplet and verified on the live URL (screenshot from a Windows machine — never run a browser on the droplet), with a short summary of exactly what changed.

Think through the best approach first, then complete the whole task autonomously before stopping.

You are working on ONE specific website: stumblicious.com (live at https://stumblicious.com/), a Web Curation site in the WholeTech network, served from /var/www/stumblicious.com/ on the droplet (ssh root@143.198.182.180). Every instruction below is for stumblicious.com and no other site — make all changes on stumblicious.com. Act like a senior editor and web engineer who knows the Web Curation niche cold.

GOAL (on stumblicious.com)
Publish: A 'what changed in 2026' round-up for Web Curation — on the site stumblicious.com.

WHY IT MATTERS
Trending angle in Web Curation. A fresh, data-backed page on this draws search and gives the site a current spine. Link it from the homepage and /sources/.

HOW TO APPROACH IT
1. Research the angle right now using at least 3 authoritative, current sources; capture the real, specific data points.
2. Write a focused page: a clear thesis, the evidence, what it concretely means for a Web Curation reader, and a clear next step.
3. Cite and link each source inline and date the piece so its freshness is obvious.
4. Link it from the homepage and /sources/, and submit it via IndexNow.

CONSTRAINTS
- Match the site's existing design, fonts, layout and voice — never a generic AI-template look.
- Keep the HTML valid; do not break nav, internal links, the promo bar, analytics, or ad units.
- Use only real, current, accurately-attributed information — no fabricated facts, numbers, or sources.
- Reuse the site's existing Google Analytics and AdSense.

WHAT "DONE" LOOKS LIKE
- A reader learns something current and genuinely useful they could not get from a thin AI page.
- At least 3 real sources are cited and linked; no fabricated numbers.
- The page is internally linked and submitted for fast indexing.
- Deployed to the droplet and verified on the live URL (screenshot from a Windows machine — never run a browser on the droplet), with a short summary of exactly what changed.

Think through the best approach first, then complete the whole task autonomously before stopping.

You are working on ONE specific website: stumblicious.com (live at https://stumblicious.com/), a Web Curation site in the WholeTech network, served from /var/www/stumblicious.com/ on the droplet (ssh root@143.198.182.180). Every instruction below is for stumblicious.com and no other site — make all changes on stumblicious.com. Act like a senior editor and web engineer who knows the Web Curation niche cold.

GOAL (on stumblicious.com)
Submit-a-site form with draft-only notifications — on the site stumblicious.com.

WHY IT MATTERS
Add a submission form for readers to nominate discoveries, generating Gmail DRAFT notifications for review — never auto-sending. Submissions turn readers into contributors, keep the stumble pipeline full without scraping, and every accepted nominee becomes a site owner with a genuine reason to link and share their feature. (Fable enhanced action — also on the global worksheet.)

HOW TO APPROACH IT
1. Plan the cleanest way to accomplish this, then execute it well.
2. Use real, current information and match the existing quality of the site.

CONSTRAINTS
- Match the site's existing design, fonts, layout and voice — never a generic AI-template look.
- Keep the HTML valid; do not break nav, internal links, the promo bar, analytics, or ad units.
- Use only real, current, accurately-attributed information — no fabricated facts, numbers, or sources.
- Reuse the site's existing Google Analytics and AdSense.

WHAT "DONE" LOOKS LIKE
- The task is fully done, accurate, and live.
- Deployed to the droplet and verified on the live URL (screenshot from a Windows machine — never run a browser on the droplet), with a short summary of exactly what changed.

Think through the best approach first, then complete the whole task autonomously before stopping.

You are working on ONE specific website: stumblicious.com (live at https://stumblicious.com/), a Web Curation site in the WholeTech network, served from /var/www/stumblicious.com/ on the droplet (ssh root@143.198.182.180). Every instruction below is for stumblicious.com and no other site — make all changes on stumblicious.com. Act like a senior editor and web engineer who knows the Web Curation niche cold.

GOAL (on stumblicious.com)
Single AdSense unit between stumble cards — on the site stumblicious.com.

WHY IT MATTERS
Place one pub-7759195213529699 unit after the third card on daily pages and one on directory pages, capped there deliberately. A discovery site's entire value is that clicks lead somewhere delightful; restrained placement earns steady RPM from browsing sessions without making stumblicious feel like the clickbait it curates against. (Fable enhanced action — also on the global worksheet.)

HOW TO APPROACH IT
1. Plan the cleanest way to accomplish this, then execute it well.
2. Use real, current information and match the existing quality of the site.

CONSTRAINTS
- Match the site's existing design, fonts, layout and voice — never a generic AI-template look.
- Keep the HTML valid; do not break nav, internal links, the promo bar, analytics, or ad units.
- Use only real, current, accurately-attributed information — no fabricated facts, numbers, or sources.
- Reuse the site's existing Google Analytics and AdSense.

WHAT "DONE" LOOKS LIKE
- The task is fully done, accurate, and live.
- Deployed to the droplet and verified on the live URL (screenshot from a Windows machine — never run a browser on the droplet), with a short summary of exactly what changed.

Think through the best approach first, then complete the whole task autonomously before stopping.

You are working on ONE specific website: stumblicious.com (live at https://stumblicious.com/), a Web Curation site in the WholeTech network, served from /var/www/stumblicious.com/ on the droplet (ssh root@143.198.182.180). Every instruction below is for stumblicious.com and no other site — make all changes on stumblicious.com. Act like a senior editor and web engineer who knows the Web Curation niche cold.

GOAL (on stumblicious.com)
Deepen the main Web Curation guide with current data and examples — on the site stumblicious.com.

WHY IT MATTERS
Audit the flagship page; replace anything stale, add concrete numbers and a clear next step. Depth is what separates this from a thin template.

HOW TO APPROACH IT
1. Audit the flagship Web Curation page and list exactly what is stale, thin, or vague.
2. Replace stale material with current data, add concrete real examples, tighten the structure, and add a clear next step for the reader.
3. Cross-link to /sources/ and the most related pages.

CONSTRAINTS
- Match the site's existing design, fonts, layout and voice — never a generic AI-template look.
- Keep the HTML valid; do not break nav, internal links, the promo bar, analytics, or ad units.
- Use only real, current, accurately-attributed information — no fabricated facts, numbers, or sources.
- Reuse the site's existing Google Analytics and AdSense.

WHAT "DONE" LOOKS LIKE
- Depth and accuracy match the best competitor page in Web Curation.
- No filler and no fabricated specifics — every number is current and sourced.
- Deployed to the droplet and verified on the live URL (screenshot from a Windows machine — never run a browser on the droplet), with a short summary of exactly what changed.

Think through the best approach first, then complete the whole task autonomously before stopping.

You are working on ONE specific website: stumblicious.com (live at https://stumblicious.com/), a Web Curation site in the WholeTech network, served from /var/www/stumblicious.com/ on the droplet (ssh root@143.198.182.180). Every instruction below is for stumblicious.com and no other site — make all changes on stumblicious.com. Act like a senior editor and web engineer who knows the Web Curation niche cold.

GOAL (on stumblicious.com)
Desk rabbit-holes gadget shelf, Amazon tag — on the site stumblicious.com.

WHY IT MATTERS
Curate a colinfirthfan-20 page of genuinely delightful physical curiosities — fidget mechanisms, odd desk toys, puzzle boxes, strange stationery — framed as 'the offline stumble.' It matches the audience's novelty appetite honestly, and curiosity-driven browsers convert well on low-cost impulse items they didn't know existed. (Fable enhanced action — also on the global worksheet.)

HOW TO APPROACH IT
1. Plan the cleanest way to accomplish this, then execute it well.
2. Use real, current information and match the existing quality of the site.

CONSTRAINTS
- Match the site's existing design, fonts, layout and voice — never a generic AI-template look.
- Keep the HTML valid; do not break nav, internal links, the promo bar, analytics, or ad units.
- Use only real, current, accurately-attributed information — no fabricated facts, numbers, or sources.
- Reuse the site's existing Google Analytics and AdSense.

WHAT "DONE" LOOKS LIKE
- The task is fully done, accurate, and live.
- Deployed to the droplet and verified on the live URL (screenshot from a Windows machine — never run a browser on the droplet), with a short summary of exactly what changed.

Think through the best approach first, then complete the whole task autonomously before stopping.

You are working on ONE specific website: stumblicious.com (live at https://stumblicious.com/), a Web Curation site in the WholeTech network, served from /var/www/stumblicious.com/ on the droplet (ssh root@143.198.182.180). Every instruction below is for stumblicious.com and no other site — make all changes on stumblicious.com. Act like a senior editor and web engineer who knows the Web Curation niche cold.

GOAL (on stumblicious.com)
Weekly digest newsletter as the owned asset — on the site stumblicious.com.

WHY IT MATTERS
Embed wholetech.com/newsletter/widget.js offering a Friday 'seven best stumbles' digest. A curation site's audience is inherently drive-by; the list converts drive-bys into a recurring readership that compounds, and a proven open rate is what makes any future honest sponsorship conversation possible. (Fable enhanced action — also on the global worksheet.)

HOW TO APPROACH IT
1. Plan the cleanest way to accomplish this, then execute it well.
2. Use real, current information and match the existing quality of the site.

CONSTRAINTS
- Match the site's existing design, fonts, layout and voice — never a generic AI-template look.
- Keep the HTML valid; do not break nav, internal links, the promo bar, analytics, or ad units.
- Use only real, current, accurately-attributed information — no fabricated facts, numbers, or sources.
- Reuse the site's existing Google Analytics and AdSense.

WHAT "DONE" LOOKS LIKE
- The task is fully done, accurate, and live.
- Deployed to the droplet and verified on the live URL (screenshot from a Windows machine — never run a browser on the droplet), with a short summary of exactly what changed.

Think through the best approach first, then complete the whole task autonomously before stopping.

You are working on ONE specific website: stumblicious.com (live at https://stumblicious.com/), a Web Curation site in the WholeTech network, served from /var/www/stumblicious.com/ on the droplet (ssh root@143.198.182.180). Every instruction below is for stumblicious.com and no other site — make all changes on stumblicious.com. Act like a senior editor and web engineer who knows the Web Curation niche cold.

GOAL (on stumblicious.com)
Slot-machine serendipity meets modern grid — on the site stumblicious.com.

WHY IT MATTERS
Art-direct around the pull-the-lever moment: an oversized die-and-lever Stumble button with a satisfying roll animation, Geocities-adjacent color pops and marquee-era accents disciplined inside a clean modern card grid. The design should celebrate early-web chaos affectionately while staying legible — nostalgia as seasoning, not costume. (Fable enhanced action — also on the global worksheet.)

HOW TO APPROACH IT
1. Snapshot the current design first (save a dated copy under the site) so there is a rollback point.
2. Apply the change so the site looks distinctive and intentional — never a generic template — then run the Agents-First scorer.
3. Verify on the live render, not on a file:// preview.

CONSTRAINTS
- Match the site's existing design, fonts, layout and voice — never a generic AI-template look.
- Keep the HTML valid; do not break nav, internal links, the promo bar, analytics, or ad units.
- Use only real, current, accurately-attributed information — no fabricated facts, numbers, or sources.
- Reuse the site's existing Google Analytics and AdSense.

WHAT "DONE" LOOKS LIKE
- The Agents-First score is 90+ and the page looks designed rather than defaulted, with a rollback point saved.
- Deployed to the droplet and verified on the live URL (screenshot from a Windows machine — never run a browser on the droplet), with a short summary of exactly what changed.

Think through the best approach first, then complete the whole task autonomously before stopping.

You are working on ONE specific website: stumblicious.com (live at https://stumblicious.com/), a Web Curation site in the WholeTech network, served from /var/www/stumblicious.com/ on the droplet (ssh root@143.198.182.180). Every instruction below is for stumblicious.com and no other site — make all changes on stumblicious.com. Act like a senior editor and web engineer who knows the Web Curation niche cold.

GOAL (on stumblicious.com)
Add an FAQ answering the top questions in Web Curation — on the site stumblicious.com.

WHY IT MATTERS
Pull the real questions people ask in Web Curation and answer them plainly. Mark up with FAQ schema for rich results.

HOW TO APPROACH IT
1. Audit the flagship Web Curation page and list exactly what is stale, thin, or vague.
2. Replace stale material with current data, add concrete real examples, tighten the structure, and add a clear next step for the reader.
3. Cross-link to /sources/ and the most related pages.

CONSTRAINTS
- Match the site's existing design, fonts, layout and voice — never a generic AI-template look.
- Keep the HTML valid; do not break nav, internal links, the promo bar, analytics, or ad units.
- Use only real, current, accurately-attributed information — no fabricated facts, numbers, or sources.
- Reuse the site's existing Google Analytics and AdSense.

WHAT "DONE" LOOKS LIKE
- Depth and accuracy match the best competitor page in Web Curation.
- No filler and no fabricated specifics — every number is current and sourced.
- Deployed to the droplet and verified on the live URL (screenshot from a Windows machine — never run a browser on the droplet), with a short summary of exactly what changed.

Think through the best approach first, then complete the whole task autonomously before stopping.

You are working on ONE specific website: stumblicious.com (live at https://stumblicious.com/), a Web Curation site in the WholeTech network, served from /var/www/stumblicious.com/ on the droplet (ssh root@143.198.182.180). Every instruction below is for stumblicious.com and no other site — make all changes on stumblicious.com. Act like a senior editor and web engineer who knows the Web Curation niche cold.

GOAL (on stumblicious.com)
ItemList schema on daily stumble pages — on the site stumblicious.com.

WHY IT MATTERS
Mark each daily five-link page with ItemList schema naming and describing every pick, plus a site-level FAQ ('what is stumblicious', 'how are sites chosen'). Keep both sitemaps fresh with the daily cadence, ping IndexNow per edition, and hold af-deploy.py at 90+ so agents can traverse the archive. (Fable enhanced action — also on the global worksheet.)

HOW TO APPROACH IT
1. Plan the cleanest way to accomplish this, then execute it well.
2. Use real, current information and match the existing quality of the site.

CONSTRAINTS
- Match the site's existing design, fonts, layout and voice — never a generic AI-template look.
- Keep the HTML valid; do not break nav, internal links, the promo bar, analytics, or ad units.
- Use only real, current, accurately-attributed information — no fabricated facts, numbers, or sources.
- Reuse the site's existing Google Analytics and AdSense.

WHAT "DONE" LOOKS LIKE
- The task is fully done, accurate, and live.
- Deployed to the droplet and verified on the live URL (screenshot from a Windows machine — never run a browser on the droplet), with a short summary of exactly what changed.

Think through the best approach first, then complete the whole task autonomously before stopping.

You are working on ONE specific website: stumblicious.com (live at https://stumblicious.com/), a Web Curation site in the WholeTech network, served from /var/www/stumblicious.com/ on the droplet (ssh root@143.198.182.180). Every instruction below is for stumblicious.com and no other site — make all changes on stumblicious.com. Act like a senior editor and web engineer who knows the Web Curation niche cold.

GOAL (on stumblicious.com)
Card-shuffle animation on every stumble — on the site stumblicious.com.

WHY IT MATTERS
When the Stumble button fires, animate the result as a card dealt off a shuffled deck — slight rotation, drop shadow, snap into place — with the mood chip stamped on like a suit. The physical deal makes randomness feel generous rather than arbitrary, and the moment is inherently screen-recordable for shares. (Fable enhanced action — also on the global worksheet.)

HOW TO APPROACH IT
1. Snapshot the current design first (save a dated copy under the site) so there is a rollback point.
2. Apply the change so the site looks distinctive and intentional — never a generic template — then run the Agents-First scorer.
3. Verify on the live render, not on a file:// preview.

CONSTRAINTS
- Match the site's existing design, fonts, layout and voice — never a generic AI-template look.
- Keep the HTML valid; do not break nav, internal links, the promo bar, analytics, or ad units.
- Use only real, current, accurately-attributed information — no fabricated facts, numbers, or sources.
- Reuse the site's existing Google Analytics and AdSense.

WHAT "DONE" LOOKS LIKE
- The Agents-First score is 90+ and the page looks designed rather than defaulted, with a rollback point saved.
- Deployed to the droplet and verified on the live URL (screenshot from a Windows machine — never run a browser on the droplet), with a short summary of exactly what changed.

Think through the best approach first, then complete the whole task autonomously before stopping.

You are working on ONE specific website: stumblicious.com (live at https://stumblicious.com/), a Web Curation site in the WholeTech network, served from /var/www/stumblicious.com/ on the droplet (ssh root@143.198.182.180). Every instruction below is for stumblicious.com and no other site — make all changes on stumblicious.com. Act like a senior editor and web engineer who knows the Web Curation niche cold.

GOAL (on stumblicious.com)
One-click random stumble from vetted JSON — on the site stumblicious.com.

WHY IT MATTERS
Ship the core mechanic: a Stumble button pulling a random entry client-side from the vetted-sites JSON, with mood-chip filtering. This is the product the domain name promises, it requires no backend, and every press is another chance a visitor finds something worth sharing with stumblicious attached. (Fable enhanced action — also on the global worksheet.)

HOW TO APPROACH IT
1. Plan the cleanest way to accomplish this, then execute it well.
2. Use real, current information and match the existing quality of the site.

CONSTRAINTS
- Match the site's existing design, fonts, layout and voice — never a generic AI-template look.
- Keep the HTML valid; do not break nav, internal links, the promo bar, analytics, or ad units.
- Use only real, current, accurately-attributed information — no fabricated facts, numbers, or sources.
- Reuse the site's existing Google Analytics and AdSense.

WHAT "DONE" LOOKS LIKE
- The task is fully done, accurate, and live.
- Deployed to the droplet and verified on the live URL (screenshot from a Windows machine — never run a browser on the droplet), with a short summary of exactly what changed.

Think through the best approach first, then complete the whole task autonomously before stopping.

You are working on ONE specific website: stumblicious.com (live at https://stumblicious.com/), a Web Curation site in the WholeTech network, served from /var/www/stumblicious.com/ on the droplet (ssh root@143.198.182.180). Every instruction below is for stumblicious.com and no other site — make all changes on stumblicious.com. Act like a senior editor and web engineer who knows the Web Curation niche cold.

GOAL (on stumblicious.com)
Keep the live / news feed updated on cadence (2×/day) — on the site stumblicious.com.

WHY IT MATTERS
Confirm the cron is publishing fresh, attributed Web Curation items and pinging IndexNow each run.

HOW TO APPROACH IT
1. Confirm or set up the auto-publishing /live/ or /news/ feed for Web Curation: attributed, links out, and running on a 2x/day cadence.
2. Make sure each run pings IndexNow and the items are genuinely fresh and relevant.

CONSTRAINTS
- Match the site's existing design, fonts, layout and voice — never a generic AI-template look.
- Keep the HTML valid; do not break nav, internal links, the promo bar, analytics, or ad units.
- Use only real, current, accurately-attributed information — no fabricated facts, numbers, or sources.
- Reuse the site's existing Google Analytics and AdSense.

WHAT "DONE" LOOKS LIKE
- The feed publishes fresh, correctly-attributed Web Curation items on schedule and is discovered quickly.
- Deployed to the droplet and verified on the live URL (screenshot from a Windows machine — never run a browser on the droplet), with a short summary of exactly what changed.

Think through the best approach first, then complete the whole task autonomously before stopping.

You are working on ONE specific website: stumblicious.com (live at https://stumblicious.com/), a Web Curation site in the WholeTech network, served from /var/www/stumblicious.com/ on the droplet (ssh root@143.198.182.180). Every instruction below is for stumblicious.com and no other site — make all changes on stumblicious.com. Act like a senior editor and web engineer who knows the Web Curation niche cold.

GOAL (on stumblicious.com)
Apply a WT-Skins kit to lift the design and Agents-First score — on the site stumblicious.com.

WHY IT MATTERS
The page looks generic without a kit. Roll the cluster's kit (snapshot the old design first) to get the AF score into the 90s.

HOW TO APPROACH IT
1. Snapshot the current design first (save a dated copy under the site) so there is a rollback point.
2. Apply the change so the site looks distinctive and intentional — never a generic template — then run the Agents-First scorer.
3. Verify on the live render, not on a file:// preview.

CONSTRAINTS
- Match the site's existing design, fonts, layout and voice — never a generic AI-template look.
- Keep the HTML valid; do not break nav, internal links, the promo bar, analytics, or ad units.
- Use only real, current, accurately-attributed information — no fabricated facts, numbers, or sources.
- Reuse the site's existing Google Analytics and AdSense.

WHAT "DONE" LOOKS LIKE
- The Agents-First score is 90+ and the page looks designed rather than defaulted, with a rollback point saved.
- Deployed to the droplet and verified on the live URL (screenshot from a Windows machine — never run a browser on the droplet), with a short summary of exactly what changed.

Think through the best approach first, then complete the whole task autonomously before stopping.

You are working on ONE specific website: stumblicious.com (live at https://stumblicious.com/), a Web Curation site in the WholeTech network, served from /var/www/stumblicious.com/ on the droplet (ssh root@143.198.182.180). Every instruction below is for stumblicious.com and no other site — make all changes on stumblicious.com. Act like a senior editor and web engineer who knows the Web Curation niche cold.

GOAL (on stumblicious.com)
Submit updated URLs via IndexNow and refresh the sitemap — on the site stumblicious.com.

WHY IT MATTERS
Ping IndexNow for new and changed pages so they're discovered fast; confirm the sitemap lists /act/, /sources/, /videos/.

HOW TO APPROACH IT
1. Implement the specific change (structured data / IndexNow / internal links) precisely and completely.
2. Validate it — rich-results test for schema, confirm the IndexNow submission, or check the internal-link graph.

CONSTRAINTS
- Match the site's existing design, fonts, layout and voice — never a generic AI-template look.
- Keep the HTML valid; do not break nav, internal links, the promo bar, analytics, or ad units.
- Use only real, current, accurately-attributed information — no fabricated facts, numbers, or sources.
- Reuse the site's existing Google Analytics and AdSense.

WHAT "DONE" LOOKS LIKE
- The change validates cleanly and the affected pages are eligible for richer results or faster discovery.
- Deployed to the droplet and verified on the live URL (screenshot from a Windows machine — never run a browser on the droplet), with a short summary of exactly what changed.

Think through the best approach first, then complete the whole task autonomously before stopping.

You are working on ONE specific website: stumblicious.com (live at https://stumblicious.com/), a Web Curation site in the WholeTech network, served from /var/www/stumblicious.com/ on the droplet (ssh root@143.198.182.180). Every instruction below is for stumblicious.com and no other site — make all changes on stumblicious.com. Act like a senior editor and web engineer who knows the Web Curation niche cold.

GOAL (on stumblicious.com)
Add Article + FAQ structured data — on the site stumblicious.com.

WHY IT MATTERS
Schema.org markup earns richer results and helps answer-engines cite the page correctly.

HOW TO APPROACH IT
1. Implement the specific change (structured data / IndexNow / internal links) precisely and completely.
2. Validate it — rich-results test for schema, confirm the IndexNow submission, or check the internal-link graph.

CONSTRAINTS
- Match the site's existing design, fonts, layout and voice — never a generic AI-template look.
- Keep the HTML valid; do not break nav, internal links, the promo bar, analytics, or ad units.
- Use only real, current, accurately-attributed information — no fabricated facts, numbers, or sources.
- Reuse the site's existing Google Analytics and AdSense.

WHAT "DONE" LOOKS LIKE
- The change validates cleanly and the affected pages are eligible for richer results or faster discovery.
- Deployed to the droplet and verified on the live URL (screenshot from a Windows machine — never run a browser on the droplet), with a short summary of exactly what changed.

Think through the best approach first, then complete the whole task autonomously before stopping.

You are working on ONE specific website: stumblicious.com (live at https://stumblicious.com/), a Web Curation site in the WholeTech network, served from /var/www/stumblicious.com/ on the droplet (ssh root@143.198.182.180). Every instruction below is for stumblicious.com and no other site — make all changes on stumblicious.com. Act like a senior editor and web engineer who knows the Web Curation niche cold.

GOAL (on stumblicious.com)
Strengthen internal links between the core, /sources/ and this worksheet — on the site stumblicious.com.

WHY IT MATTERS
A tight internal-link web spreads authority and keeps visitors moving through the site.

HOW TO APPROACH IT
1. Implement the specific change (structured data / IndexNow / internal links) precisely and completely.
2. Validate it — rich-results test for schema, confirm the IndexNow submission, or check the internal-link graph.

CONSTRAINTS
- Match the site's existing design, fonts, layout and voice — never a generic AI-template look.
- Keep the HTML valid; do not break nav, internal links, the promo bar, analytics, or ad units.
- Use only real, current, accurately-attributed information — no fabricated facts, numbers, or sources.
- Reuse the site's existing Google Analytics and AdSense.

WHAT "DONE" LOOKS LIKE
- The change validates cleanly and the affected pages are eligible for richer results or faster discovery.
- Deployed to the droplet and verified on the live URL (screenshot from a Windows machine — never run a browser on the droplet), with a short summary of exactly what changed.

Think through the best approach first, then complete the whole task autonomously before stopping.

You are working on ONE specific website: stumblicious.com (live at https://stumblicious.com/), a Web Curation site in the WholeTech network, served from /var/www/stumblicious.com/ on the droplet (ssh root@143.198.182.180). Every instruction below is for stumblicious.com and no other site — make all changes on stumblicious.com. Act like a senior editor and web engineer who knows the Web Curation niche cold.

GOAL (on stumblicious.com)
Refresh the /videos/ playlist with current, embeddable uploads — on the site stumblicious.com.

WHY IT MATTERS
Swap any dead or stale videos; keep only authoritative, non-clickbait Web Curation videos. Re-verify embeds.

HOW TO APPROACH IT
1. Find 6-10 authoritative, genuinely useful Web Curation videos that ALLOW embedding — verify each embed actually plays.
2. Curate them on /videos/ with short context, and drop anything clickbait, get-rich-quick, or low-substance.

CONSTRAINTS
- Match the site's existing design, fonts, layout and voice — never a generic AI-template look.
- Keep the HTML valid; do not break nav, internal links, the promo bar, analytics, or ad units.
- Use only real, current, accurately-attributed information — no fabricated facts, numbers, or sources.
- Reuse the site's existing Google Analytics and AdSense.

WHAT "DONE" LOOKS LIKE
- Every embed plays, the set is authoritative and on-topic, and the page reuses the site's GA + AdSense.
- Deployed to the droplet and verified on the live URL (screenshot from a Windows machine — never run a browser on the droplet), with a short summary of exactly what changed.

Think through the best approach first, then complete the whole task autonomously before stopping.

You are working on ONE specific website: stumblicious.com (live at https://stumblicious.com/), a Web Curation site in the WholeTech network, served from /var/www/stumblicious.com/ on the droplet (ssh root@143.198.182.180). Every instruction below is for stumblicious.com and no other site — make all changes on stumblicious.com. Act like a senior editor and web engineer who knows the Web Curation niche cold.

GOAL (on stumblicious.com)
Verify affiliate links and ad placement resolve correctly — on the site stumblicious.com.

WHY IT MATTERS
Spot-check Skimlinks / Amazon links and AdSense units; remove anything broken or low-value.

HOW TO APPROACH IT
1. Make the change — verify affiliate links resolve, or add the shared affiliate + AdSense setup.
2. Keep commerce low-key and well-integrated: subtle in-content links, not a bolted-on banner block.

CONSTRAINTS
- Match the site's existing design, fonts, layout and voice — never a generic AI-template look.
- Keep the HTML valid; do not break nav, internal links, the promo bar, analytics, or ad units.
- Use only real, current, accurately-attributed information — no fabricated facts, numbers, or sources.
- Reuse the site's existing Google Analytics and AdSense.

WHAT "DONE" LOOKS LIKE
- Links resolve and can earn, and placement is tasteful and does not hurt the reading experience.
- Deployed to the droplet and verified on the live URL (screenshot from a Windows machine — never run a browser on the droplet), with a short summary of exactly what changed.

Think through the best approach first, then complete the whole task autonomously before stopping.