M skills/pi-browser-harness/SKILL.md → skills/pi-browser-harness/SKILL.md

@@ -41,6 +41,8 @@ Pass `@(x,y)` from `browser_snapshot` straight to `browser_click`. No screenshot round-trip.
 
 `browser_web_search` and `browser_read_page` each run in their own isolated tab and never touch the user's current tab. Pair them for a research question: search for candidate URLs, then read the promising ones.
 
+`browser_read_page` runs a reader-mode heuristic: JS-rendered pages (X, social SPAs, login walls) can come back as a short nav shell — if the result is tiny or reads like chrome, don't treat it as the page; re-read the `og:title`/`og:description` meta or pull the content with `browser_execute_js` (e.g. `document.querySelector('article')?.innerText`).
+
 ## Connection Setup
 
 Browser control is **on-demand** — the daemon does NOT start automatically.

M src/domains/readpage/read-page.ts → src/domains/readpage/read-page.ts

@@ -158,7 +158,7 @@ "Read a page as clean article text — main content with nav/ads/boilerplate stripped. Pass a url (opened in an isolated tab, read, then closed) or a targetId of an owned tab. Reader-mode counterpart to browser_web_search.",
   promptSnippet: "Read a page's main content as clean text",
   promptGuidelines: [
     "Returns readable main-article text — use this over browser_snapshot/browser_execute_js when you want an article's content for reading or research.",
-    "Boilerplate-heavy or structure-less pages fall back to bounded body text rather than erroring.",
+    "If the result is short or reads like a nav shell (e.g. JS-rendered social pages like X, login-walled pages), extraction likely ran before the app rendered or missed the content — do NOT treat the shell as the page. Re-read via og:title/og:description, or use browser_execute_js with an article/content selector.",
   ],
   parameters: ReadPageArgs,
   concurrency: "parallel",