From 4b7a3f45b7f8d68f6192d77fff0c5a403b2c0920 Mon Sep 17 00:00:00 2001
From: Mohamed Saleh Zaied <eng.mohamedszaied@gmail.com>
Date: Sun, 28 Jun 2026 20:44:51 +0300
Subject: [PATCH 1/4] Add youtube/upload domain skill (battle-tested Studio
 flow)

Step-by-step browser-harness flow for publishing to YouTube via Studio: Create->Upload videos, file via DOM.setFileInputFiles (last input), contenteditable #textbox title/desc, 'Made for Kids' curly-quote gotcha, coordinate-click visibility radios, Publish confirmation. Captured 2026-06-28 while publishing the Skilly Builders demo.
---
 domain-skills/youtube/upload.md | 51 +++++++++++++++++++++++++++++++++
 1 file changed, 51 insertions(+)
 create mode 100644 domain-skills/youtube/upload.md

diff --git a/domain-skills/youtube/upload.md b/domain-skills/youtube/upload.md
new file mode 100644
index 00000000..b4b111b5
--- /dev/null
+++ b/domain-skills/youtube/upload.md
@@ -0,0 +1,51 @@
+# youtube/upload — publish a video to YouTube via Studio (battle-tested 2026-06-28)
+
+Driving `studio.youtube.com` with browser-harness against the user's logged-in Chrome. This is the durable map of the upload flow — follow it and the first try works. The whole flow is ~5 steps but several controls resist JS selectors and need **coordinate clicks** (noted below).
+
+## Pre-flight
+- User must be logged into the target channel in Chrome. Confirm: `page_info()` on `studio.youtube.com` returns `.../channel/<CHANNEL_ID>` and title contains "YouTube Studio".
+- **Check for an existing video first.** The channel dashboard "Latest video performance" card shows the most recent upload — don't create a duplicate. To reuse an existing video's URL instead of uploading: go to `.../videos/upload`, scrape `a[href*="/video/"]` → `/video/(<id>)/` and build `https://youtu.be/<id>`.
+- Have the absolute path to the local `.mp4` ready.
+
+## The flow
+
+### 1. Open the upload dialog
+- Click **Create** (top-right): match `aria-label === "Create"` (an `<button>`/`ytcp-button`). Get rect, `click(cx,cy)`.
+- Menu appears with: Create post / **Upload videos** / Go live. Click the element whose trimmed innerText is exactly `Upload videos` (`children.length<=1`, width>0).
+
+### 2. Set the file (this triggers the upload immediately)
+- After the dialog opens, find file inputs via CDP and set the **last** one:
+  ```python
+  doc=cdp("DOM.getDocument", depth=-1, pierce=True)
+  nodes=cdp("DOM.querySelectorAll", nodeId=doc["root"]["nodeId"], selector='input[type=file]')["nodeIds"]
+  cdp("DOM.setFileInputFiles", files=[ABS_MP4_PATH], nodeId=nodes[-1])
+  ```
+- Wait ~8s. The dialog shows "Upload complete … Processing will begin shortly". The **public URL is available immediately** in the "Video link" anchor: scrape `a[href*="youtu.be"]` → e.g. `https://youtu.be/qm3aVsWDARY`. Capture it now.
+
+### 3. Details — title + description
+- Title and description are **contenteditable `#textbox` divs** (NOT inputs). There are two large ones; sort by `y`: first = Title (pre-filled from filename), second = Description.
+- Replace the title: `click()` it → `document.execCommand('selectAll',false,null)` → `cdp("Input.insertText", text=TITLE)`.
+- Description: `click()` it → `cdp("Input.insertText", text=DESC)`.
+- **Trap:** external links in the description are NOT clickable until the channel completes a one-off verification ("To make external links clickable, first complete a one-off verification"). The link text still shows; fine for a backlink mention.
+
+### 4. Audience — "Made for Kids" (REQUIRED, or Next is blocked)
+- Radios are `tp-yt-paper-radio-button`. The "no" option's text is literally **`No, it's not 'Made for Kids'`** (curly quotes around Made for Kids — do NOT match `/not made for kids/`; match `/no,/i` AND `/made for kids/i`).
+- `scrollIntoView({block:'center'})` then `click` its left edge (`rect.x+18, rect.y+height/2`). Verify `aria-checked === "true"`.
+
+### 5. Advance + Visibility + Publish
+- Click **Next** (trimmed innerText `Next`, width>0) **3×** with ~3s waits: Details → Video elements → Checks → Visibility. Confirm you're on Visibility (`document.body.innerText` contains "Save or publish").
+- **Visibility radios resist JS selectors — use COORDINATE clicks.** The three options stack: Private, Unlisted, **Public** (3rd). Selecting any visibility enables the bottom-right button and relabels it `Save`→`Publish`.
+  - If a JS query for the radio returns null, click by coordinate. (See coordinate-conversion note below.)
+  - Verify selection worked by checking that a `Publish`/`Save` button with `aria-disabled!=="true"` now exists near the bottom (`rect.y>700`).
+- Click **Publish**. Success = a **"Video published"** dialog appears with the share link (`youtu.be/<id>`) + Promote button.
+
+## Gotchas (field-tested)
+- **Visibility radios: coordinate clicks only.** `[role=radio]`/`tp-yt-paper-radio-button` text matches fail because the label/description live in sibling nodes. Locate the Public row from a screenshot and click it.
+- **Coordinate conversion:** screenshots come back at 2× retina (e.g. 3840 wide shown at 2000). For `click(x,y)` (CSS px on a ~1920 viewport): `css = displayed_screenshot_coord × 0.96`.
+- **Clicking Publish/Save with no visibility chosen is safely blocked** — a "You need to choose a visibility setting" tooltip shows and the Visibility step turns red; nothing publishes. So a mis-fire here is harmless; just select Public and retry.
+- **Auto-save:** the dialog header reads "Saved as private" / "Saving…" throughout — the video exists as private from the moment of upload; publishing flips it to the chosen visibility.
+- **Don't re-upload an existing video** — check the dashboard card first.
+
+## Content best-practices (for the human filling fields)
+- **Title** ≤100 chars, keyword-first. **Description**: 1–2 line hook + links (with timestamps if long). **Tags** in Show More. **Custom thumbnail** beats auto-generated for CTR. Add **end screens / cards** in the "Video elements" step if you skipped it.
+- For embedding in an article (HackerNoon, etc.): paste the `youtu.be/<id>` URL on its own line — most editors auto-embed the player. Public or Unlisted both embed.

From afa309ded6a1b2cdab714ec4464453b9aa037ca6 Mon Sep 17 00:00:00 2001
From: Mohamed Saleh Zaied <eng.mohamedszaied@gmail.com>
Date: Sun, 28 Jun 2026 22:14:59 +0300
Subject: [PATCH 2/4] Add devto/publish domain skill (instant dofollow backlink
 flow)

DEV.to markdown-editor publish flow: front-matter+body in one textarea via React-safe native setter, published flag for draft/live, Save-changes-publishes, and the 'Invalid authenticity token' CSRF gotcha (reload to refresh). Captured 2026-06-28 cross-posting the Skilly Realtime-API article.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
---
 domain-skills/devto/publish.md | 51 ++++++++++++++++++++++++++++++++++
 1 file changed, 51 insertions(+)
 create mode 100644 domain-skills/devto/publish.md

diff --git a/domain-skills/devto/publish.md b/domain-skills/devto/publish.md
new file mode 100644
index 00000000..a1597c03
--- /dev/null
+++ b/domain-skills/devto/publish.md
@@ -0,0 +1,51 @@
+# devto/publish — publish an article to DEV.to via the markdown editor (battle-tested 2026-06-28)
+
+Driving `dev.to/new` with browser-harness against the user's logged-in Chrome. DEV.to gives an **instant, no-queue, high-DR dofollow backlink** in the article body — ideal for cross-posting an existing technical article. Far simpler than rich editors: it's one markdown textarea with Jekyll front matter.
+
+## Pre-flight
+- User logged into dev.to in Chrome. `new_tab("https://dev.to/new")` → if the editor (`textarea#article_body_markdown`) is present you're in; if you see a login wall, pause and ask.
+- The "basic markdown editor" puts **everything in one textarea** — title/tags/description go in YAML front matter at the top, body below.
+
+## The flow
+
+### 1. Build the content (front matter + markdown body)
+```
+---
+title: <Title, plain text>
+published: true            # true = publish now; false = save as draft
+description: <≤ ~150 chars>
+tags: programming, ai, webdev, saas    # comma-separated, MAX 4, lowercase, no spaces/hyphens inside a tag
+canonical_url: <optional — set to the original if cross-posting, to avoid duplicate-content>
+---
+
+<markdown body>
+```
+- **YouTube / Tweet embeds:** liquid tag on its own line — `{% embed https://youtu.be/<id> %}`.
+- Body links are normal markdown `[text](https://...)` and render **dofollow**.
+
+### 2. Insert it (React-safe native setter — don't type)
+The editor is one big controlled textarea. Set the value via the native setter + fire input/change (typing char-by-char is slow and flaky):
+```python
+md = open("/path/post.md").read()
+js("""(function(){
+  var t=document.querySelector('textarea#article_body_markdown')||
+        [].slice.call(document.querySelectorAll('textarea')).sort((a,b)=>b.getBoundingClientRect().height-a.getBoundingClientRect().height)[0];
+  var setter=Object.getOwnPropertyDescriptor(window.HTMLTextAreaElement.prototype,'value').set;
+  setter.call(t, %s); t.dispatchEvent(new Event('input',{bubbles:true})); t.dispatchEvent(new Event('change',{bubbles:true}));
+  return 'set:'+t.value.length;
+})()""" % json.dumps(md))
+```
+
+### 3. Publish
+- Scroll to the bottom; the button is **"Save changes"** (there is no separate "Publish" button in this editor). With `published: true` in the front matter, clicking **Save changes publishes**; with `published: false` it saves a draft.
+- Success = the URL changes to `https://dev.to/<user>/<slug>`.
+
+## Gotchas (field-tested)
+- **"Invalid authenticity token" (CSRF) on save** — the #1 failure. The Rails CSRF token goes stale if the page sat open a while (or across a failed submit). **Fix: `goto("https://dev.to/new")` to reload (fresh token), re-insert the content, save again.** Don't dwell between loading the editor and submitting.
+- **No "Publish" button** — it's "Save changes"; the `published:` front-matter flag decides draft vs live. Don't hunt for a Publish button.
+- **Tags:** max 4, lowercase, each a single token (e.g. `webdev`, not `web-dev`); invalid tags can block the save.
+- **Cover image** is optional but boosts the home-feed; drag-drop only (skip in automation, the user can add later).
+- Cross-posting the same article to multiple sites (HackerNoon, Hashnode, dev.to): set `canonical_url` to the original on the secondary copies if you care about duplicate-content; for pure-backlink plays it's optional.
+
+## Why it's worth it
+Instant publish, no moderator queue, DEV.to is high domain authority, and body links are dofollow — so a single cross-post of an existing article = one of the "5 backlinks that matter," in ~3 minutes. Hashnode works the same way (markdown editor + instant publish).

From 322f3e7b630f2647246e412aa2fb2a60c708ce35 Mon Sep 17 00:00:00 2001
From: Mohamed Saleh Zaied <eng.mohamedszaied@gmail.com>
Date: Sun, 28 Jun 2026 22:34:35 +0300
Subject: [PATCH 3/4] Address cubic review: rename devto->dev-to, quote YAML,
 import json, match video by title

Fixes flagged by the automated reviewer (cubic) on PR #476: (1) move devto/publish.md into the existing dev-to/ dir (+ cross-ref scraping.md); (2) quote YAML front-matter scalars to survive colons/# in title/description; (3) add missing 'import json' to the snippet; (4) youtube/upload existing-video check now matches by title across the full Content list instead of trusting the latest dashboard card.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
---
 domain-skills/{devto => dev-to}/publish.md | 12 ++++++++----
 domain-skills/youtube/upload.md            |  2 +-
 2 files changed, 9 insertions(+), 5 deletions(-)
 rename domain-skills/{devto => dev-to}/publish.md (83%)

diff --git a/domain-skills/devto/publish.md b/domain-skills/dev-to/publish.md
similarity index 83%
rename from domain-skills/devto/publish.md
rename to domain-skills/dev-to/publish.md
index a1597c03..d1465c35 100644
--- a/domain-skills/devto/publish.md
+++ b/domain-skills/dev-to/publish.md
@@ -1,7 +1,9 @@
-# devto/publish — publish an article to DEV.to via the markdown editor (battle-tested 2026-06-28)
+# dev-to/publish — publish an article to DEV.to via the markdown editor (battle-tested 2026-06-28)
 
 Driving `dev.to/new` with browser-harness against the user's logged-in Chrome. DEV.to gives an **instant, no-queue, high-DR dofollow backlink** in the article body — ideal for cross-posting an existing technical article. Far simpler than rich editors: it's one markdown textarea with Jekyll front matter.
 
+> Companion skill: `dev-to/scraping.md` (reading dev.to). This file covers writing/publishing.
+
 ## Pre-flight
 - User logged into dev.to in Chrome. `new_tab("https://dev.to/new")` → if the editor (`textarea#article_body_markdown`) is present you're in; if you see a login wall, pause and ask.
 - The "basic markdown editor" puts **everything in one textarea** — title/tags/description go in YAML front matter at the top, body below.
@@ -11,21 +13,23 @@ Driving `dev.to/new` with browser-harness against the user's logged-in Chrome. D
 ### 1. Build the content (front matter + markdown body)
 ```
 ---
-title: <Title, plain text>
+title: "<Title — ALWAYS quote it: a colon or # in the title breaks YAML otherwise>"
 published: true            # true = publish now; false = save as draft
-description: <≤ ~150 chars>
+description: "<≤ ~150 chars — quote it too (colons/# are common in descriptions)>"
 tags: programming, ai, webdev, saas    # comma-separated, MAX 4, lowercase, no spaces/hyphens inside a tag
-canonical_url: <optional — set to the original if cross-posting, to avoid duplicate-content>
+canonical_url: "<optional original URL, if cross-posting, to avoid duplicate-content>"
 ---
 
 <markdown body>
 ```
+Quote any user-supplied scalar (title/description/canonical_url). dev.to's front matter is YAML, so an unquoted value containing `:` or a leading `#` will fail the parse or be misread, and the publish silently breaks.
 - **YouTube / Tweet embeds:** liquid tag on its own line — `{% embed https://youtu.be/<id> %}`.
 - Body links are normal markdown `[text](https://...)` and render **dofollow**.
 
 ### 2. Insert it (React-safe native setter — don't type)
 The editor is one big controlled textarea. Set the value via the native setter + fire input/change (typing char-by-char is slow and flaky):
 ```python
+import json
 md = open("/path/post.md").read()
 js("""(function(){
   var t=document.querySelector('textarea#article_body_markdown')||
diff --git a/domain-skills/youtube/upload.md b/domain-skills/youtube/upload.md
index b4b111b5..2a98ca7d 100644
--- a/domain-skills/youtube/upload.md
+++ b/domain-skills/youtube/upload.md
@@ -4,7 +4,7 @@ Driving `studio.youtube.com` with browser-harness against the user's logged-in C
 
 ## Pre-flight
 - User must be logged into the target channel in Chrome. Confirm: `page_info()` on `studio.youtube.com` returns `.../channel/<CHANNEL_ID>` and title contains "YouTube Studio".
-- **Check for an existing video first.** The channel dashboard "Latest video performance" card shows the most recent upload — don't create a duplicate. To reuse an existing video's URL instead of uploading: go to `.../videos/upload`, scrape `a[href*="/video/"]` → `/video/(<id>)/` and build `https://youtu.be/<id>`.
+- **Check for an existing video first (match by title, not position).** Go to `.../videos/upload` (the full Content list, not just the dashboard's "Latest video" card — that card only shows the most recent one and will miss older duplicates). Scrape each row's title cell **and** its `a[href*="/video/"]` → `/video/(<id>)/` together, then find the row whose title matches the video you intend to publish; build `https://youtu.be/<id>` from that row's id. Never assume the first/most-recent row is the match — confirm the title, or you'll reuse (or skip over) the wrong video.
 - Have the absolute path to the local `.mp4` ready.
 
 ## The flow

From 6f99dc816fa6212aa701235f6125d3506b28aa12 Mon Sep 17 00:00:00 2001
From: Mohamed Saleh Zaied <eng.mohamedszaied@gmail.com>
Date: Mon, 29 Jun 2026 01:49:09 +0300
Subject: [PATCH 4/4] Add google-search-console/check domain skill (indexing +
 sitemap flow)
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Browser-harness flow for GSC: sitemap submit (domain-property full-URL gotcha + two-input targeting), Pages indexing report, URL inspection, Performance — plus how to read 'Discovered – currently not indexed' (authority/crawl-budget, not a bug). Captured 2026-06-29 submitting the tryskilly.app sitemap + reading the indexing report.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
---
 domain-skills/google-search-console/check.md | 42 ++++++++++++++++++++
 1 file changed, 42 insertions(+)
 create mode 100644 domain-skills/google-search-console/check.md

diff --git a/domain-skills/google-search-console/check.md b/domain-skills/google-search-console/check.md
new file mode 100644
index 00000000..43d6a1d6
--- /dev/null
+++ b/domain-skills/google-search-console/check.md
@@ -0,0 +1,42 @@
+# google-search-console/check — inspect indexing & submit sitemaps (battle-tested 2026-06-29)
+
+Driving `search.google.com/search-console` with browser-harness against the user's logged-in Google account. **No API here** unless service-account/OAuth is set up — this is the browser path. GSC is a heavy Angular SPA: use coordinate clicks + `Input.insertText`, and **read reports via screenshots** (text extraction is sparse).
+
+## Property URL
+Everything is scoped by `resource_id`. For a **domain property** it's `sc-domain:<domain>` (e.g. `sc-domain:tryskilly.app`). Deep-link directly:
+- Overview: `…/search-console?resource_id=sc-domain:<domain>`
+- Sitemaps: `…/search-console/sitemaps?resource_id=sc-domain:<domain>`
+- Pages (indexing): `…/search-console/index?resource_id=sc-domain:<domain>`
+- Performance: `…/search-console/performance/search_analytics?resource_id=sc-domain:<domain>`
+
+Confirm login + property by reading the top-left property name and the left nav (Overview / Performance / URL inspection / Pages / Sitemaps…). If it redirects to a Google login, pause and ask the user.
+
+## The 4 checks (priority order)
+
+### 1. Submit / re-submit the sitemap (highest value after a deploy)
+- Go to the Sitemaps page. The "Add a new sitemap" card has an input + **SUBMIT** button.
+- **GOTCHA — domain properties require the FULL URL.** Entering a relative path (`sitemap-index.xml`) returns **"Invalid sitemap address"**. Enter `https://<domain>/sitemap-index.xml`.
+- **GOTCHA — two inputs on the page.** The top header has an "Inspect any URL" bar AND the card has the sitemap input — both are `<input>`. `document.querySelector('input')` grabs the **wrong** (top) one. Target the sitemap card's input (it sits lower in the main content, ~y 239 at 1080p), then `click` it → `Input.insertText` the full URL → click SUBMIT.
+- Success = a green **"Sitemap submitted successfully"** dialog; the submitted-sitemaps table shows the row with Status **Success**, Last read, and **Discovered pages**.
+- Re-submitting after adding pages nudges a fresh read (Google also re-reads on its own schedule).
+
+### 2. Pages (indexing) report — the real "is everything indexed?" answer
+- Two tiles: **Indexed** (count) and **Not indexed (N reasons)**.
+- Scroll to **"Why pages aren't indexed"** for the per-reason table.
+- **GOTCHA — data lags ~2 weeks** ("Last update" date at top). Pages deployed today won't show here for a while; the sitemap submit (step 1) is what accelerates discovery.
+
+### 3. URL inspection
+- Top "Inspect any URL" bar → type a full URL → Enter. Shows whether it's on Google + why. Click **"Request indexing"** to push a priority page into the queue (rate-limited, ~10–20/day).
+
+### 4. Performance
+- Impressions / clicks / avg position / top queries. Early-stage sites show little — that's expected.
+
+## Interpreting "not indexed" reasons
+- **"Discovered – currently not indexed"** → Google found the URL (via sitemap/links) but deprioritized crawling it. This is a **domain-authority / crawl-budget** signal, NOT a technical error. Fix: earn backlinks, strengthen internal linking (hub→spoke), Request-Indexing the key pages, and wait. If most pages sit here, **do not mass-generate more pages** — they'll pile up unindexed; raise authority first.
+- **"Crawled – currently not indexed"** → crawled but judged low-value/thin/duplicate. Improve content quality/uniqueness.
+- **"Page with redirect"** → usually benign (trailing-slash or http→https canonicalization); the redirect target is the indexed URL.
+- **"Duplicate without user-selected canonical" / "Alternate page with canonical"** → set/confirm canonical tags.
+- **"Not found (404)"** → a referenced URL is broken; find + fix or remove the reference.
+
+## Why no API
+GSC has a Search Console API (sitemaps, URL inspection, search analytics) but it needs a **service account added to the property** or an **OAuth client** — neither is set up here (a prior attempt stalled on service-account propagation). Until that's configured, this browser flow is the way. If you do set it up, the `searchconsole.googleapis.com` URL-inspection + `webmasters` sitemaps endpoints replace steps 1–3.