@koda-sl/baker-cli 0.439.0 → 0.440.0

package/README.md

@@ -6714,6 +6714,7 @@ baker landing inspiration code <section id>     # the standalone bundle
 baker landing inspiration page <source id>      # a whole page as a section sequence
 baker landing inspiration sequences "b2b saas pricing page" --scope all   # what follows what, across many pages
 baker landing inspiration add https://linear.app --note "client likes this density"
+baker landing inspiration add https://linear.app --refresh   # read a saved page again — the site changed
 baker landing inspiration favorites             # what this company has saved
 baker landing inspiration retry-failed          # study again every saved page that failed
 baker landing inspiration scrape <url> --out .baker/<dir>   # capture any page now, synchronously
@@ -6729,6 +6730,7 @@ baker landing inspiration scrape <url> --out .baker/<dir>   # capture any page n
 - **Compact by default, `--full` when you have chosen something.** Every list command returns what you need to *pick* a row; the 600-character "why it works" paragraph, the classification facets and the markup ride behind `--full` on `search`, `view`, `page`, `favorites` and `code`. A default `favorites` used to cost more than a default search for a question — "what does this client keep saving?" — that the facets alone answer.
 - **`sequences` answers the ordering question search structurally can't.** Search ranks individual sections and caps each site at three, so a page's top-to-bottom order never survives its results, and `page` recovers exactly one page's order. `sequences` counts adjacency across many pages at once and returns `openers`, `transitions` (`{ from, to, count }`, most frequent first) and `closers`, filtered by `--scope --domain --section-type --limit`. It returns page **ids only** — `page` stays the one way to read a single page — and always reports `pages_considered` alongside `pages_returned`, because a frequency over an unstated denominator is not evidence. `--full` counts over every matching page and prints the whole table.
 - **`page_id` on every search row** is what `page <id>` takes, so "how does this page sequence its sections?" is reachable from a result rather than only right after `add`. `favorites --limit` is capped at 100.
+- **Pages are re-read, on request and on a schedule.** A page is a copy taken when it was read, so `add --refresh` reads it again when the live site has changed (refused while a read is under way or within 10 minutes of the last one), and the library re-reads pages older than 30 days on its own, a small batch a day. A re-read that fails before it starts replacing the copy keeps the previous one and is reported as `refresh_failed_at`. `page` reports `read_at` and hints at `--refresh` (with the address the page is filed under) when the copy is over 30 days old.
 - **`add` studies in the background; `scrape` returns with the page on disk.** `add` grows the shared library and takes minutes, so it can never answer "build our page like this one" within the same turn. `scrape` runs the identical capture locally and blocks until it finishes, writing section screenshots, standalone markup, a whole-page reproduction and `report.html`. `--no-motion` is the single biggest lever on runtime; `--no-mobile`, `--no-code` and `--no-report` skip further passes.
 - **Filming is the expensive pass, and it runs three sections at a time.** Every moving section is filmed in its own fresh page load — the only way to catch an entrance animation before it fires — so a page with ten moving sections pays for ten full loads. Measured on one heavy page, filming was 74% of the capture. The takes are independent, so they overlap; a page that cannot be captured inside the library's budget now says so instead of timing out silently.
 - **A page we can't open says why.** A dead certificate, an address that doesn't resolve, a site that won't answer or one that is simply too slow each produce their own plain-language reason rather than a browser error.

package/dist/cli.js

@@ -100940,6 +100940,12 @@ function fidelityHint(fidelity) {
   }
   return `Reproduction fidelity ${fidelity.toFixed(2)} \u2014 this section is scroll- or JS-driven and the extracted markup does NOT render like the original. Use the screenshot and the notes; treat the code as a hint only.`;
 }
+var STALE_COPY_AFTER_MS = 30 * 24 * 60 * 60 * 1e3;
+function staleCopyHint(studiedAt, filedUrl, now = Date.now()) {
+  if (!studiedAt || now - studiedAt < STALE_COPY_AFTER_MS) return null;
+  const date = new Date(studiedAt).toLocaleDateString("en-US", { month: "short", day: "numeric", timeZone: "UTC" });
+  return `This copy of the page was read on ${date}. If the user says the live page looks different, run: baker landing inspiration add ${filedUrl} --refresh`;
+}
 function reportError2(error) {
   if (error instanceof ApiError) {
     writeJson({ ok: false, error: { code: error.code, message: error.message } });
@@ -101117,37 +101123,67 @@ registerSchema({
   description: "Add someone else's landing page to the reference library and save it to this company. Returns immediately \u2014 studying a page takes a few minutes, so never wait on it. Use it when the client names a site they admire, or when you find a competitor page worth learning from. Never point it at a page we built: it is refused, because the source is already in the project and reading a screenshot of it tells you less than opening the file.",
   args: {
     url: { type: "string", description: "Landing page address", required: true },
-    note: { type: "string", description: "Why this page is worth keeping", required: false }
+    note: { type: "string", description: "Why this page is worth keeping", required: false },
+    refresh: {
+      type: "boolean",
+      description: "Study the page again even if it is already in the library. Use when the user says the page has changed or the saved sections look outdated",
+      required: false
+    }
   }
 });
+function addHints(data, { refresh }) {
+  if (data.refreshDeclined === "studying") {
+    return [
+      "This page is already being studied \u2014 no second study was started. Search for it later in the turn or in a later one."
+    ];
+  }
+  if (data.refreshDeclined === "too_soon") {
+    return [
+      "This page was just studied a few minutes ago, so it was not studied again \u2014 its sections are already current.",
+      `Search it with: baker landing inspiration search --domain ${new URL(data.canonicalUrl).hostname}`
+    ];
+  }
+  if (data.alreadyKnown) {
+    return [
+      "This page was already in the library, so nothing was re-studied \u2014 its sections are searchable now.",
+      `Search it with: baker landing inspiration search --domain ${new URL(data.canonicalUrl).hostname}`,
+      "If the user says the live page looks different from what the library shows, run this again with --refresh."
+    ];
+  }
+  return [
+    refresh ? "Studying it again from the live site \u2014 this takes a few minutes. Until it lands, search results still show the old version, so don't build from them yet." : "Studying this page takes a few minutes. Don't wait on it \u2014 carry on, and search for it later in the turn or in a later one.",
+    "Meanwhile, search what is already in the library: baker landing inspiration search '<what you want to see>' --scope all",
+    "This is saved and being studied now, so it is listed as already applied \u2014 tell the user it is kept whatever they do with this chat.",
+    // The save is not a promise that the study succeeded. An agent that
+    // assumes it did will later search its own favorites, find nothing,
+    // and have no idea why.
+    "The save sticks even if the page turns out to be unreadable \u2014 `baker landing inspiration favorites` reports those separately as unreadable, so check there rather than assuming this page is searchable."
+  ];
+}
 var addCommand = defineCommand221({
   meta: {
     name: "add",
-    description: "Add someone else's landing page to the reference library. Example: baker landing inspiration add https://linear.app --note 'the client likes this density'"
+    description: "Add someone else's landing page to the reference library. Example: baker landing inspiration add https://linear.app --note 'the client likes this density'. Add --refresh to study a saved page again after the site changed"
   },
   args: {
     url: { type: "positional", description: "Landing page address", required: true },
-    note: { type: "string", description: "Why this page is worth keeping", required: false }
+    note: { type: "string", description: "Why this page is worth keeping", required: false },
+    refresh: {
+      type: "boolean",
+      description: "Study it again even if it is already in the library",
+      required: false,
+      default: false
+    }
   },
   run: async ({ args: args8 }) => {
     try {
       const data = await apiPost("/api/landing-inspiration/add", {
         url: args8.url,
         note: args8.note,
-        favorite: true
+        favorite: true,
+        refresh: args8.refresh || void 0
       });
-      const hints2 = data.alreadyKnown ? [
-        "This page was already in the library, so nothing was re-studied \u2014 its sections are searchable now.",
-        `Search it with: baker landing inspiration search --domain ${new URL(data.canonicalUrl).hostname}`
-      ] : [
-        "Studying this page takes a few minutes. Don't wait on it \u2014 carry on, and search for it later in the turn or in a later one.",
-        "Meanwhile, search what is already in the library: baker landing inspiration search '<what you want to see>' --scope all",
-        "This is saved and being studied now, so it is listed as already applied \u2014 tell the user it is kept whatever they do with this chat.",
-        // The save is not a promise that the study succeeded. An agent that
-        // assumes it did will later search its own favorites, find nothing,
-        // and have no idea why.
-        "The save sticks even if the page turns out to be unreadable \u2014 `baker landing inspiration favorites` reports those separately as unreadable, so check there rather than assuming this page is searchable."
-      ];
+      const hints2 = addHints(data, { refresh: args8.refresh });
       writeJson({
         ok: true,
         data: { id: data.sourceId, url: data.canonicalUrl, status: data.status, already_known: data.alreadyKnown },
@@ -101421,6 +101457,8 @@ var pageCommand2 = defineCommand224({
           archetype: data.archetype,
           stack: data.detectedStack,
           page_fidelity: data.pageFidelity,
+          read_at: data.studiedAt ? new Date(data.studiedAt).toISOString().slice(0, 10) : null,
+          ...data.refreshFailedAt ? { refresh_failed_at: new Date(data.refreshFailedAt).toISOString().slice(0, 10) } : {},
           blueprint: data.sectionOrder,
           sections: data.sections.map((section2) => ({
             id: section2.id,
@@ -101441,7 +101479,11 @@ var pageCommand2 = defineCommand224({
           "The blueprint is the section order top to bottom \u2014 the answer to 'how do good pages in this category sequence themselves'.",
           "Use `baker landing inspiration view <section id>` for any section worth a closer look.",
           ...full ? [] : ["Add --full for the idea behind every section, once you know which page you care about."],
-          INSPIRATION_HINTS.adapt
+          INSPIRATION_HINTS.adapt,
+          ...[staleCopyHint(data.studiedAt, data.filedUrl)].filter((hint) => hint !== null),
+          ...data.refreshFailedAt ? [
+            "The last attempt to read this page again failed, so this is the earlier copy. Tell the user if the live page may differ."
+          ] : []
         ]
       });
     } catch (error) {
@@ -104166,7 +104208,7 @@ Subcommands:
   baker landing inspiration code <id>         \u2014 its standalone HTML+CSS, for structure only
   baker landing inspiration page <id>         \u2014 a whole page as a sequence: how it orders its sections
   baker landing inspiration sequences "<q>"   \u2014 what follows what across MANY pages: openers, transitions, closers
-  baker landing inspiration add <url>         \u2014 add a page to the library and save it to this company (returns immediately)
+  baker landing inspiration add <url>         \u2014 add a page to the library and save it to this company (returns immediately); --refresh re-studies a page that changed
   baker landing inspiration favorites         \u2014 what this company has saved; the default search scope
   baker landing inspiration favorite <id>     \u2014 save a section
   baker landing inspiration unfavorite <id>   \u2014 unsave a section