@pseolint/core 0.3.2 → 0.4.1

package/dist/page-filter.js

@@ -0,0 +1,86 @@
+/**
+ * Page-skip detection helpers used by the auditor pipeline (v0.4.1).
+ *
+ * Two policy filters live here, each gated by an AuditOption:
+ *   - `detectNoindex(page)` honours `<meta name="robots" content="noindex">`
+ *     and `X-Robots-Tag: noindex` HTTP headers. The site owner already told
+ *     Google not to index these pages — auditing them produces noise the
+ *     reader can't act on.
+ *   - `detectAuthPage(page)` heuristically classifies pages as
+ *     login / signup / password-reset based on three signals (password input
+ *     in a thin body, title matches the auth regex, H1 matches the auth
+ *     regex). Two signals are required for a positive verdict, which keeps
+ *     the false-positive rate low: a marketing landing page with a single
+ *     password input or a single auth-shaped heading won't trip it.
+ *
+ * Both functions are pure and synchronous so they can be called inside the
+ * `parsedPages` map step without disrupting the existing pipeline shape.
+ */
+/**
+ * Returns true when the page is explicitly noindex'd via either the parsed
+ * `<meta name="robots">` content or the `X-Robots-Tag` HTTP response header.
+ * Match is case-insensitive and substring-based to tolerate combined
+ * directives like `"noindex, nofollow"` or `"index, noindex"` (the latter is
+ * a common bug — `noindex` wins per spec).
+ */
+export function detectNoindex(page) {
+    const metaRobots = (page.robotsMeta ?? "").toLowerCase();
+    const xRobots = (page.httpMeta?.xRobotsTag ?? "").toLowerCase();
+    return metaRobots.includes("noindex") || xRobots.includes("noindex");
+}
+/**
+ * Auth-shaped page title pattern, applied AFTER stripping the brand suffix.
+ * Matches "Sign in", "Sign-In", "Log in", "Log-Out", "Register", "Sign up",
+ * "Forgot password", "Reset password", "Create account", "Verify email" and
+ * common variants. Case-insensitive; whole-string match (anchored).
+ */
+const AUTH_TITLE_REGEX = /^(sign[-\s]?in|sign[-\s]?up|sign[-\s]?out|log[-\s]?in|log[-\s]?out|register|forgot[-\s]password|reset[-\s]password|create[-\s](?:an[-\s])?account|verify[-\s](?:your[-\s])?email|two[-\s]?factor|account[-\s]recovery)$/i;
+/**
+ * Common brand-suffix separators. We split on the FIRST occurrence and keep
+ * everything before it — typical pattern is `Sign in | MyApp` or
+ * `Sign in - MyApp` or `Sign in · MyApp`. Without this strip, the regex
+ * would never match because the title would be `Sign in | MyApp`, not just
+ * `Sign in`.
+ */
+const BRAND_SEPARATORS = [" | ", " - ", " — ", " · ", " : ", " :: "];
+function stripBrandSuffix(title) {
+    let cut = title.length;
+    for (const sep of BRAND_SEPARATORS) {
+        const idx = title.indexOf(sep);
+        if (idx >= 0 && idx < cut)
+            cut = idx;
+    }
+    return title.slice(0, cut).trim();
+}
+/**
+ * Heuristic auth-page detection. Returns `isAuth: true` when 2+ signals
+ * fire (high confidence). Single-signal pages return `isAuth: false` but
+ * still expose the signal list for diagnostics — useful for tracking
+ * borderline cases that might warrant a manual config exclude.
+ */
+export function detectAuthPage(page) {
+    const signals = [];
+    const wordCount = (page.contentText ?? "").split(/\s+/).filter(Boolean).length;
+    const h2Count = (page.headings?.h2 ?? []).length;
+    const hasPasswordInput = /<input\b[^>]*\btype\s*=\s*["']?password["']?/i.test(page.html ?? "");
+    if (hasPasswordInput && wordCount < 200 && h2Count < 3) {
+        signals.push("password-input");
+    }
+    const cleanedTitle = stripBrandSuffix(page.title ?? "");
+    if (cleanedTitle && AUTH_TITLE_REGEX.test(cleanedTitle)) {
+        signals.push("auth-title");
+    }
+    const firstH1 = (page.headings?.h1 ?? [])[0]?.trim() ?? "";
+    if (firstH1 && AUTH_TITLE_REGEX.test(firstH1)) {
+        signals.push("auth-h1");
+    }
+    return { isAuth: signals.length >= 2, signals };
+}
+export function pageSkipReason(page, options) {
+    if (options.respectNoindex && detectNoindex(page))
+        return "noindex";
+    if (options.skipDetectedAuth && detectAuthPage(page).isAuth)
+        return "auth-detected";
+    return null;
+}
+//# sourceMappingURL=page-filter.js.map

package/dist/page-filter.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"page-filter.js","sourceRoot":"","sources":["../src/page-filter.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;GAiBG;AAIH;;;;;;GAMG;AACH,MAAM,UAAU,aAAa,CAAC,IAAgB;IAC5C,MAAM,UAAU,GAAG,CAAC,IAAI,CAAC,UAAU,IAAI,EAAE,CAAC,CAAC,WAAW,EAAE,CAAC;IACzD,MAAM,OAAO,GAAG,CAAC,IAAI,CAAC,QAAQ,EAAE,UAAU,IAAI,EAAE,CAAC,CAAC,WAAW,EAAE,CAAC;IAChE,OAAO,UAAU,CAAC,QAAQ,CAAC,SAAS,CAAC,IAAI,OAAO,CAAC,QAAQ,CAAC,SAAS,CAAC,CAAC;AACvE,CAAC;AAED;;;;;GAKG;AACH,MAAM,gBAAgB,GACpB,0NAA0N,CAAC;AAE7N;;;;;;GAMG;AACH,MAAM,gBAAgB,GAAG,CAAC,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,MAAM,CAAC,CAAC;AAErE,SAAS,gBAAgB,CAAC,KAAa;IACrC,IAAI,GAAG,GAAG,KAAK,CAAC,MAAM,CAAC;IACvB,KAAK,MAAM,GAAG,IAAI,gBAAgB,EAAE,CAAC;QACnC,MAAM,GAAG,GAAG,KAAK,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC;QAC/B,IAAI,GAAG,IAAI,CAAC,IAAI,GAAG,GAAG,GAAG;YAAE,GAAG,GAAG,GAAG,CAAC;IACvC,CAAC;IACD,OAAO,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;AACpC,CAAC;AAOD;;;;;GAKG;AACH,MAAM,UAAU,cAAc,CAAC,IAAgB;IAC7C,MAAM,OAAO,GAAmC,EAAE,CAAC;IAEnD,MAAM,SAAS,GAAG,CAAC,IAAI,CAAC,WAAW,IAAI,EAAE,CAAC,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,MAAM,CAAC;IAC/E,MAAM,OAAO,GAAG,CAAC,IAAI,CAAC,QAAQ,EAAE,EAAE,IAAI,EAAE,CAAC,CAAC,MAAM,CAAC;IACjD,MAAM,gBAAgB,GAAG,+CAA+C,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,IAAI,EAAE,CAAC,CAAC;IAC/F,IAAI,gBAAgB,IAAI,SAAS,GAAG,GAAG,IAAI,OAAO,GAAG,CAAC,EAAE,CAAC;QACvD,OAAO,CAAC,IAAI,CAAC,gBAAgB,CAAC,CAAC;IACjC,CAAC;IAED,MAAM,YAAY,GAAG,gBAAgB,CAAC,IAAI,CAAC,KAAK,IAAI,EAAE,CAAC,CAAC;IACxD,IAAI,YAAY,IAAI,gBAAgB,CAAC,IAAI,CAAC,YAAY,CAAC,EAAE,CAAC;QACxD,OAAO,CAAC,IAAI,CAAC,YAAY,CAAC,CAAC;IAC7B,CAAC;IAED,MAAM,OAAO,GAAG,CAAC,IAAI,CAAC,QAAQ,EAAE,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,IAAI,EAAE,CAAC;IAC3D,IAAI,OAAO,IAAI,gBAAgB,CAAC,IAAI,CAAC,OAAO,CAAC,EAAE,CAAC;QAC9C,OAAO,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;IAC1B,CAAC;IAED,OAAO,EAAE,MAAM,EAAE,OAAO,CAAC,MAAM,IAAI,CAAC,EAAE,OAAO,EAAE,CAAC;AAClD,CAAC;AAaD,MAAM,UAAU,cAAc,CAC5B,IAAgB,EAChB,OAAwB;IAExB,IAAI,OAAO,CAAC,cAAc,IAAI,aAAa,CAAC,IAAI,CAAC;QAAE,OAAO,SAAS,CAAC;IACpE,IAAI,OAAO,CAAC,gBAAgB,IAAI,cAAc,CAAC,IAAI,CAAC,CAAC,MAAM;QAAE,OAAO,eAAe,CAAC;IACpF,OAAO,IAAI,CAAC;AACd,CAAC"}

package/dist/rule-references.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"rule-references.d.ts","sourceRoot":"","sources":["../src/rule-references.ts"],"names":[],"mappings":"AAAA,eAAO,MAAM,eAAe,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CA6ClD,CAAC"}
+{"version":3,"file":"rule-references.d.ts","sourceRoot":"","sources":["../src/rule-references.ts"],"names":[],"mappings":"AAAA,eAAO,MAAM,eAAe,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAuClD,CAAC"}

package/dist/rule-references.js

@@ -9,13 +9,11 @@ export const RULE_REFERENCES = {
     "spam/template-coverage": "https://developers.google.com/search/docs/essentials/spam-policies#doorway-pages",
     "content/unique-value": "https://developers.google.com/search/docs/fundamentals/creating-helpful-content",
     "content/meta-uniqueness": "https://developers.google.com/search/docs/appearance/snippet#meta-descriptions",
-    "content/heading-uniqueness": "https://developers.google.com/search/docs/appearance/snippet#headings",
     "content/missing-author": "https://developers.google.com/search/docs/fundamentals/creating-helpful-content#eeat",
     "content/eeat-signals": "https://developers.google.com/search/docs/fundamentals/creating-helpful-content#eeat",
     "links/orphan-pages": "https://developers.google.com/search/docs/crawling-indexing/links-crawlable",
     "links/dead-ends": "https://developers.google.com/search/docs/crawling-indexing/links-crawlable",
     "links/cluster-connectivity": "https://developers.google.com/search/docs/crawling-indexing/links-crawlable",
-    "links/hub-pages": "https://developers.google.com/search/docs/crawling-indexing/links-crawlable",
     "links/link-depth": "https://developers.google.com/search/docs/crawling-indexing/links-crawlable",
     "tech/robots-compliance": "https://developers.google.com/search/docs/crawling-indexing/robots/intro",
     "tech/canonical-consistency": "https://developers.google.com/search/docs/crawling-indexing/consolidate-duplicate-urls",
@@ -24,13 +22,10 @@ export const RULE_REFERENCES = {
     "tech/sitemap-completeness": "https://developers.google.com/search/docs/crawling-indexing/sitemaps/overview",
     "tech/redirect-chain": "https://developers.google.com/search/docs/crawling-indexing/301-redirects",
     "tech/soft-404": "https://developers.google.com/search/docs/crawling-indexing/soft-404-errors",
-    "tech/og-completeness": "https://developers.google.com/search/docs/appearance/snippet",
     "tech/hreflang-consistency": "https://developers.google.com/search/docs/specialty/international/managing-multi-regional-sites",
     "schema/json-ld-valid": "https://developers.google.com/search/docs/appearance/structured-data/intro-structured-data",
     "schema/required-fields": "https://developers.google.com/search/docs/appearance/structured-data/intro-structured-data",
     "schema/consistency": "https://developers.google.com/search/docs/appearance/structured-data/intro-structured-data",
-    "cannibal/title-overlap": "https://developers.google.com/search/docs/crawling-indexing/consolidate-duplicate-urls",
-    "cannibal/keyword-collision": "https://developers.google.com/search/docs/crawling-indexing/consolidate-duplicate-urls",
     "cannibal/url-pattern": "https://developers.google.com/search/docs/crawling-indexing/consolidate-duplicate-urls",
     "data/missing-binding": "https://developers.google.com/search/docs/essentials/spam-policies#thin-content",
     "data/identical-across-pages": "https://developers.google.com/search/docs/essentials/spam-policies#scraped-content",
@@ -40,7 +35,6 @@ export const RULE_REFERENCES = {
     "aeo/faq-coverage": "https://developers.google.com/search/docs/appearance/structured-data/faqpage",
     "aeo/answer-first": "https://developers.google.com/search/docs/appearance/featured-snippets",
     "aeo/citable-facts": "https://developers.google.com/search/docs/fundamentals/creating-helpful-content",
-    "aeo/non-replicable-value": "https://developers.google.com/search/docs/fundamentals/creating-helpful-content",
     "aeo/content-modularity": "https://developers.google.com/search/docs/appearance/structured-data/intro-structured-data",
     "aeo/summary-bait": "https://developers.google.com/search/docs/fundamentals/creating-helpful-content",
 };

package/dist/rule-references.js.map

@@ -1 +1 @@
-{"version":3,"file":"rule-references.js","sourceRoot":"","sources":["../src/rule-references.ts"],"names":[],"mappings":"AAAA,MAAM,CAAC,MAAM,eAAe,GAA2B;IACrD,qBAAqB,EAAE,oFAAoF;IAC3G,kBAAkB,EAAE,kFAAkF;IACtG,mBAAmB,EAAE,iFAAiF;IACtG,wBAAwB,EAAE,iFAAiF;IAC3G,yBAAyB,EAAE,kFAAkF;IAC7G,2BAA2B,EAAE,kGAAkG;IAC/H,sBAAsB,EAAE,kFAAkF;IAC1G,wBAAwB,EAAE,kFAAkF;IAC5G,sBAAsB,EAAE,iFAAiF;IACzG,yBAAyB,EAAE,gFAAgF;IAC3G,4BAA4B,EAAE,uEAAuE;IACrG,wBAAwB,EAAE,sFAAsF;IAChH,sBAAsB,EAAE,sFAAsF;IAC9G,oBAAoB,EAAE,6EAA6E;IACnG,iBAAiB,EAAE,6EAA6E;IAChG,4BAA4B,EAAE,6EAA6E;IAC3G,iBAAiB,EAAE,6EAA6E;IAChG,kBAAkB,EAAE,6EAA6E;IACjG,wBAAwB,EAAE,0EAA0E;IACpG,4BAA4B,EAAE,wFAAwF;IACtH,iCAAiC,EAAE,wFAAwF;IAC3H,8BAA8B,EAAE,4EAA4E;IAC5G,2BAA2B,EAAE,+EAA+E;IAC5G,qBAAqB,EAAE,2EAA2E;IAClG,eAAe,EAAE,6EAA6E;IAC9F,sBAAsB,EAAE,8DAA8D;IACtF,2BAA2B,EAAE,iGAAiG;IAC9H,sBAAsB,EAAE,4FAA4F;IACpH,wBAAwB,EAAE,4FAA4F;IACtH,oBAAoB,EAAE,4FAA4F;IAClH,wBAAwB,EAAE,wFAAwF;IAClH,4BAA4B,EAAE,wFAAwF;IACtH,sBAAsB,EAAE,wFAAwF;IAChH,sBAAsB,EAAE,iFAAiF;IACzG,6BAA6B,EAAE,oFAAoF;IACnH,cAAc,EAAE,qBAAqB;IACrC,oBAAoB,EAAE,0CAA0C;IAChE,uBAAuB,EAAE,wEAAwE;IACjG,kBAAkB,EAAE,8EAA8E;IAClG,kBAAkB,EAAE,wEAAwE;IAC5F,mBAAmB,EAAE,iFAAiF;IACtG,0BAA0B,EAAE,iFAAiF;IAC7G,wBAAwB,EAAE,4FAA4F;IACtH,kBAAkB,EAAE,iFAAiF;CACtG,CAAC"}
+{"version":3,"file":"rule-references.js","sourceRoot":"","sources":["../src/rule-references.ts"],"names":[],"mappings":"AAAA,MAAM,CAAC,MAAM,eAAe,GAA2B;IACrD,qBAAqB,EAAE,oFAAoF;IAC3G,kBAAkB,EAAE,kFAAkF;IACtG,mBAAmB,EAAE,iFAAiF;IACtG,wBAAwB,EAAE,iFAAiF;IAC3G,yBAAyB,EAAE,kFAAkF;IAC7G,2BAA2B,EAAE,kGAAkG;IAC/H,sBAAsB,EAAE,kFAAkF;IAC1G,wBAAwB,EAAE,kFAAkF;IAC5G,sBAAsB,EAAE,iFAAiF;IACzG,yBAAyB,EAAE,gFAAgF;IAC3G,wBAAwB,EAAE,sFAAsF;IAChH,sBAAsB,EAAE,sFAAsF;IAC9G,oBAAoB,EAAE,6EAA6E;IACnG,iBAAiB,EAAE,6EAA6E;IAChG,4BAA4B,EAAE,6EAA6E;IAC3G,kBAAkB,EAAE,6EAA6E;IACjG,wBAAwB,EAAE,0EAA0E;IACpG,4BAA4B,EAAE,wFAAwF;IACtH,iCAAiC,EAAE,wFAAwF;IAC3H,8BAA8B,EAAE,4EAA4E;IAC5G,2BAA2B,EAAE,+EAA+E;IAC5G,qBAAqB,EAAE,2EAA2E;IAClG,eAAe,EAAE,6EAA6E;IAC9F,2BAA2B,EAAE,iGAAiG;IAC9H,sBAAsB,EAAE,4FAA4F;IACpH,wBAAwB,EAAE,4FAA4F;IACtH,oBAAoB,EAAE,4FAA4F;IAClH,sBAAsB,EAAE,wFAAwF;IAChH,sBAAsB,EAAE,iFAAiF;IACzG,6BAA6B,EAAE,oFAAoF;IACnH,cAAc,EAAE,qBAAqB;IACrC,oBAAoB,EAAE,0CAA0C;IAChE,uBAAuB,EAAE,wEAAwE;IACjG,kBAAkB,EAAE,8EAA8E;IAClG,kBAAkB,EAAE,wEAAwE;IAC5F,mBAAmB,EAAE,iFAAiF;IACtG,wBAAwB,EAAE,4FAA4F;IACtH,kBAAkB,EAAE,iFAAiF;CACtG,CAAC"}

package/dist/rules/content/unique-value.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"unique-value.d.ts","sourceRoot":"","sources":["../../../src/rules/content/unique-value.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,UAAU,EAAE,UAAU,EAAE,MAAM,gBAAgB,CAAC;AAM7D,wBAAgB,eAAe,CAAC,KAAK,EAAE,UAAU,EAAE,EAAE,cAAc,EAAE,MAAM,GAAG,UAAU,EAAE,CA0BzF"}
+{"version":3,"file":"unique-value.d.ts","sourceRoot":"","sources":["../../../src/rules/content/unique-value.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,UAAU,EAAE,UAAU,EAAE,MAAM,gBAAgB,CAAC;AAM7D,wBAAgB,eAAe,CAAC,KAAK,EAAE,UAAU,EAAE,EAAE,cAAc,EAAE,MAAM,GAAG,UAAU,EAAE,CA2BzF"}

package/dist/rules/content/unique-value.js

@@ -17,6 +17,7 @@ export function uniqueValueRule(pages, minUniqueWords) {
                 ruleId: "content/unique-value",
                 severity: "error",
                 message: `${page.url} has only ${uniqueCount} unique words (min ${minUniqueWords}).`,
+                pageUrl: page.url,
                 fix: `Add ${minUniqueWords - uniqueCount} more words of content not found on any other page.`
             });
         }

package/dist/rules/content/unique-value.js.map

@@ -1 +1 @@
-{"version":3,"file":"unique-value.js","sourceRoot":"","sources":["../../../src/rules/content/unique-value.ts"],"names":[],"mappings":"AAEA,SAAS,QAAQ,CAAC,IAAY;IAC5B,OAAO,IAAI,CAAC,WAAW,EAAE,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC;AACzD,CAAC;AAED,MAAM,UAAU,eAAe,CAAC,KAAmB,EAAE,cAAsB;IACzE,MAAM,WAAW,GAAG,IAAI,GAAG,EAAkB,CAAC;IAC9C,MAAM,UAAU,GAAG,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,QAAQ,CAAC,IAAI,CAAC,WAAW,CAAC,CAAC,CAAC;IAEnE,KAAK,MAAM,MAAM,IAAI,UAAU,EAAE,CAAC;QAChC,KAAK,MAAM,KAAK,IAAI,IAAI,GAAG,CAAC,MAAM,CAAC,EAAE,CAAC;YACpC,WAAW,CAAC,GAAG,CAAC,KAAK,EAAE,CAAC,WAAW,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;QAC5D,CAAC;IACH,CAAC;IAED,MAAM,QAAQ,GAAiB,EAAE,CAAC;IAClC,KAAK,CAAC,OAAO,CAAC,CAAC,IAAI,EAAE,GAAG,EAAE,EAAE;QAC1B,MAAM,WAAW,GAAG,IAAI,GAAG,CACzB,UAAU,CAAC,GAAG,CAAC,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,WAAW,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,KAAK,CAAC,CAAC,CACvE,CAAC,IAAI,CAAC;QACP,IAAI,WAAW,GAAG,cAAc,EAAE,CAAC;YACjC,QAAQ,CAAC,IAAI,CAAC;gBACZ,MAAM,EAAE,sBAAsB;gBAC9B,QAAQ,EAAE,OAAO;gBACjB,OAAO,EAAE,GAAG,IAAI,CAAC,GAAG,aAAa,WAAW,sBAAsB,cAAc,IAAI;gBACpF,GAAG,EAAE,OAAO,cAAc,GAAG,WAAW,qDAAqD;aAC9F,CAAC,CAAC;QACL,CAAC;IACH,CAAC,CAAC,CAAC;IAEH,OAAO,QAAQ,CAAC;AAClB,CAAC"}
+{"version":3,"file":"unique-value.js","sourceRoot":"","sources":["../../../src/rules/content/unique-value.ts"],"names":[],"mappings":"AAEA,SAAS,QAAQ,CAAC,IAAY;IAC5B,OAAO,IAAI,CAAC,WAAW,EAAE,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC;AACzD,CAAC;AAED,MAAM,UAAU,eAAe,CAAC,KAAmB,EAAE,cAAsB;IACzE,MAAM,WAAW,GAAG,IAAI,GAAG,EAAkB,CAAC;IAC9C,MAAM,UAAU,GAAG,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,QAAQ,CAAC,IAAI,CAAC,WAAW,CAAC,CAAC,CAAC;IAEnE,KAAK,MAAM,MAAM,IAAI,UAAU,EAAE,CAAC;QAChC,KAAK,MAAM,KAAK,IAAI,IAAI,GAAG,CAAC,MAAM,CAAC,EAAE,CAAC;YACpC,WAAW,CAAC,GAAG,CAAC,KAAK,EAAE,CAAC,WAAW,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;QAC5D,CAAC;IACH,CAAC;IAED,MAAM,QAAQ,GAAiB,EAAE,CAAC;IAClC,KAAK,CAAC,OAAO,CAAC,CAAC,IAAI,EAAE,GAAG,EAAE,EAAE;QAC1B,MAAM,WAAW,GAAG,IAAI,GAAG,CACzB,UAAU,CAAC,GAAG,CAAC,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,WAAW,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,KAAK,CAAC,CAAC,CACvE,CAAC,IAAI,CAAC;QACP,IAAI,WAAW,GAAG,cAAc,EAAE,CAAC;YACjC,QAAQ,CAAC,IAAI,CAAC;gBACZ,MAAM,EAAE,sBAAsB;gBAC9B,QAAQ,EAAE,OAAO;gBACjB,OAAO,EAAE,GAAG,IAAI,CAAC,GAAG,aAAa,WAAW,sBAAsB,cAAc,IAAI;gBACpF,OAAO,EAAE,IAAI,CAAC,GAAG;gBACjB,GAAG,EAAE,OAAO,cAAc,GAAG,WAAW,qDAAqD;aAC9F,CAAC,CAAC;QACL,CAAC;IACH,CAAC,CAAC,CAAC;IAEH,OAAO,QAAQ,CAAC;AAClB,CAAC"}

package/dist/rules/scope.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"scope.d.ts","sourceRoot":"","sources":["../../src/rules/scope.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AACH,MAAM,MAAM,SAAS,GAAG,MAAM,GAAG,QAAQ,CAAC;AAE1C,eAAO,MAAM,UAAU,EAAE,MAAM,CAAC,MAAM,EAAE,SAAS,CAqEhD,CAAC;AAEF,4GAA4G;AAC5G,wBAAgB,mBAAmB,CAAC,MAAM,EAAE,MAAM,GAAG,OAAO,CAE3D"}
+{"version":3,"file":"scope.d.ts","sourceRoot":"","sources":["../../src/rules/scope.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AACH,MAAM,MAAM,SAAS,GAAG,MAAM,GAAG,QAAQ,CAAC;AAE1C,eAAO,MAAM,UAAU,EAAE,MAAM,CAAC,MAAM,EAAE,SAAS,CA6DhD,CAAC;AAEF,4GAA4G;AAC5G,wBAAgB,mBAAmB,CAAC,MAAM,EAAE,MAAM,GAAG,OAAO,CAE3D"}

package/dist/rules/scope.js

@@ -10,26 +10,22 @@ export const RULE_SCOPE = {
     "spam/template-coverage": "corpus",
     // content
     "content/unique-value": "corpus",
-    "content/heading-uniqueness": "corpus",
     "content/meta-uniqueness": "corpus",
     "content/missing-author": "page",
     "content/eeat-signals": "page",
-    // links (all need the global link graph)
+    // links
     "links/orphan-pages": "corpus",
     "links/dead-ends": "corpus",
     "links/cluster-connectivity": "corpus",
-    "links/hub-pages": "corpus",
     "links/link-depth": "corpus",
     "links/unreachable-from-root": "corpus",
-    "links/hub-pages-skipped": "corpus",
-    // tech (per-page except sitemap / robots / canonical-consistency which need knownUrls)
+    // tech
     "tech/canonical-consistency": "corpus",
     "tech/canonical-noindex-conflict": "page",
     "tech/robots-noindex-conflict": "corpus",
     "tech/sitemap-completeness": "corpus",
     "tech/redirect-chain": "page",
     "tech/soft-404": "page",
-    "tech/og-completeness": "page",
     "tech/hreflang-consistency": "corpus",
     "tech/robots-compliance": "corpus",
     "tech/robots-sitemap-presence": "corpus",
@@ -37,26 +33,22 @@ export const RULE_SCOPE = {
     "schema/json-ld-valid": "page",
     "schema/required-fields": "page",
     "schema/consistency": "corpus",
-    // cannibal (cross-page by definition)
-    "cannibal/title-overlap": "corpus",
-    "cannibal/keyword-collision": "corpus",
+    // cannibal — only url-pattern survives in v0.4 (title-overlap and
+    // keyword-collision dropped due to high false-positive rates; see
+    // 2026-04-29 v0.4 redesign spec §4.3).
     "cannibal/url-pattern": "corpus",
     // data binding
     "data/missing-binding": "page",
     "data/identical-across-pages": "corpus",
     // audit-internal
     "audit/duplicate-url": "corpus",
-    // AEO (Answer Engine Optimization) — forward-compatible entries for rules
-    // shipping from a parallel core-engine change. `llms-txt` and `crawler-access`
-    // are site-wide (need host config / robots surface); the rest evaluate a
-    // single page's content structure and must run in daily diff-audits.
+    // AEO (Answer Engine Optimization).
     "aeo/llms-txt": "corpus",
     "aeo/crawler-access": "corpus",
     "aeo/freshness-signals": "page",
     "aeo/faq-coverage": "page",
     "aeo/answer-first": "page",
     "aeo/citable-facts": "page",
-    "aeo/non-replicable-value": "page",
     "aeo/content-modularity": "page",
     "aeo/summary-bait": "page",
 };

package/dist/rules/scope.js.map

@@ -1 +1 @@
-{"version":3,"file":"scope.js","sourceRoot":"","sources":["../../src/rules/scope.ts"],"names":[],"mappings":"AASA,MAAM,CAAC,MAAM,UAAU,GAA8B;IACnD,OAAO;IACP,qBAAqB,EAAE,QAAQ;IAC/B,kBAAkB,EAAE,QAAQ;IAC5B,mBAAmB,EAAE,MAAM;IAC3B,wBAAwB,EAAE,QAAQ;IAClC,yBAAyB,EAAE,QAAQ;IACnC,2BAA2B,EAAE,QAAQ;IACrC,sBAAsB,EAAE,QAAQ;IAChC,wBAAwB,EAAE,QAAQ;IAElC,UAAU;IACV,sBAAsB,EAAE,QAAQ;IAChC,4BAA4B,EAAE,QAAQ;IACtC,yBAAyB,EAAE,QAAQ;IACnC,wBAAwB,EAAE,MAAM;IAChC,sBAAsB,EAAE,MAAM;IAE9B,yCAAyC;IACzC,oBAAoB,EAAE,QAAQ;IAC9B,iBAAiB,EAAE,QAAQ;IAC3B,4BAA4B,EAAE,QAAQ;IACtC,iBAAiB,EAAE,QAAQ;IAC3B,kBAAkB,EAAE,QAAQ;IAC5B,6BAA6B,EAAE,QAAQ;IACvC,yBAAyB,EAAE,QAAQ;IAEnC,uFAAuF;IACvF,4BAA4B,EAAE,QAAQ;IACtC,iCAAiC,EAAE,MAAM;IACzC,8BAA8B,EAAE,QAAQ;IACxC,2BAA2B,EAAE,QAAQ;IACrC,qBAAqB,EAAE,MAAM;IAC7B,eAAe,EAAE,MAAM;IACvB,sBAAsB,EAAE,MAAM;IAC9B,2BAA2B,EAAE,QAAQ;IACrC,wBAAwB,EAAE,QAAQ;IAClC,8BAA8B,EAAE,QAAQ;IAExC,SAAS;IACT,sBAAsB,EAAE,MAAM;IAC9B,wBAAwB,EAAE,MAAM;IAChC,oBAAoB,EAAE,QAAQ;IAE9B,sCAAsC;IACtC,wBAAwB,EAAE,QAAQ;IAClC,4BAA4B,EAAE,QAAQ;IACtC,sBAAsB,EAAE,QAAQ;IAEhC,eAAe;IACf,sBAAsB,EAAE,MAAM;IAC9B,6BAA6B,EAAE,QAAQ;IAEvC,iBAAiB;IACjB,qBAAqB,EAAE,QAAQ;IAE/B,0EAA0E;IAC1E,+EAA+E;IAC/E,yEAAyE;IACzE,qEAAqE;IACrE,cAAc,EAAE,QAAQ;IACxB,oBAAoB,EAAE,QAAQ;IAC9B,uBAAuB,EAAE,MAAM;IAC/B,kBAAkB,EAAE,MAAM;IAC1B,kBAAkB,EAAE,MAAM;IAC1B,mBAAmB,EAAE,MAAM;IAC3B,0BAA0B,EAAE,MAAM;IAClC,wBAAwB,EAAE,MAAM;IAChC,kBAAkB,EAAE,MAAM;CAC3B,CAAC;AAEF,4GAA4G;AAC5G,MAAM,UAAU,mBAAmB,CAAC,MAAc;IAChD,OAAO,CAAC,UAAU,CAAC,MAAM,CAAC,IAAI,QAAQ,CAAC,KAAK,MAAM,CAAC;AACrD,CAAC"}
+{"version":3,"file":"scope.js","sourceRoot":"","sources":["../../src/rules/scope.ts"],"names":[],"mappings":"AASA,MAAM,CAAC,MAAM,UAAU,GAA8B;IACnD,OAAO;IACP,qBAAqB,EAAE,QAAQ;IAC/B,kBAAkB,EAAE,QAAQ;IAC5B,mBAAmB,EAAE,MAAM;IAC3B,wBAAwB,EAAE,QAAQ;IAClC,yBAAyB,EAAE,QAAQ;IACnC,2BAA2B,EAAE,QAAQ;IACrC,sBAAsB,EAAE,QAAQ;IAChC,wBAAwB,EAAE,QAAQ;IAElC,UAAU;IACV,sBAAsB,EAAE,QAAQ;IAChC,yBAAyB,EAAE,QAAQ;IACnC,wBAAwB,EAAE,MAAM;IAChC,sBAAsB,EAAE,MAAM;IAE9B,QAAQ;IACR,oBAAoB,EAAE,QAAQ;IAC9B,iBAAiB,EAAE,QAAQ;IAC3B,4BAA4B,EAAE,QAAQ;IACtC,kBAAkB,EAAE,QAAQ;IAC5B,6BAA6B,EAAE,QAAQ;IAEvC,OAAO;IACP,4BAA4B,EAAE,QAAQ;IACtC,iCAAiC,EAAE,MAAM;IACzC,8BAA8B,EAAE,QAAQ;IACxC,2BAA2B,EAAE,QAAQ;IACrC,qBAAqB,EAAE,MAAM;IAC7B,eAAe,EAAE,MAAM;IACvB,2BAA2B,EAAE,QAAQ;IACrC,wBAAwB,EAAE,QAAQ;IAClC,8BAA8B,EAAE,QAAQ;IAExC,SAAS;IACT,sBAAsB,EAAE,MAAM;IAC9B,wBAAwB,EAAE,MAAM;IAChC,oBAAoB,EAAE,QAAQ;IAE9B,kEAAkE;IAClE,kEAAkE;IAClE,uCAAuC;IACvC,sBAAsB,EAAE,QAAQ;IAEhC,eAAe;IACf,sBAAsB,EAAE,MAAM;IAC9B,6BAA6B,EAAE,QAAQ;IAEvC,iBAAiB;IACjB,qBAAqB,EAAE,QAAQ;IAE/B,oCAAoC;IACpC,cAAc,EAAE,QAAQ;IACxB,oBAAoB,EAAE,QAAQ;IAC9B,uBAAuB,EAAE,MAAM;IAC/B,kBAAkB,EAAE,MAAM;IAC1B,kBAAkB,EAAE,MAAM;IAC1B,mBAAmB,EAAE,MAAM;IAC3B,wBAAwB,EAAE,MAAM;IAChC,kBAAkB,EAAE,MAAM;CAC3B,CAAC;AAEF,4GAA4G;AAC5G,MAAM,UAAU,mBAAmB,CAAC,MAAc;IAChD,OAAO,CAAC,UAAU,CAAC,MAAM,CAAC,IAAI,QAAQ,CAAC,KAAK,MAAM,CAAC;AACrD,CAAC"}

package/dist/rules/tech/robots-sitemap-presence.d.ts

@@ -6,7 +6,15 @@ export declare function robotsSitemapPresenceRule(source: string): Promise<RuleR
  */
 /** Parse `Crawl-delay: N` from the `User-agent: *` block. Returns seconds, or 0 if unset. */
 export declare function parseCrawlDelaySeconds(robotsTxt: string): number;
-export declare function parseDisallowPatterns(robotsTxt: string): string[];
+/**
+ * Parse Disallow patterns for the given user-agents. Merges the `User-agent: *`
+ * block with any named UA blocks; a hostile site that specifically disallows
+ * `pseolint` without a wildcard still gets honored.
+ *
+ * The default UA list (`["*"]`) preserves the pre-v0.3.2 behavior for callers
+ * that don't opt into UA-specific matching.
+ */
+export declare function parseDisallowPatterns(robotsTxt: string, userAgents?: readonly string[]): string[];
 /**
  * Returns true if `urlPath` (e.g. "/blog/post") is blocked by a single Disallow pattern.
  *

package/dist/rules/tech/robots-sitemap-presence.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"robots-sitemap-presence.d.ts","sourceRoot":"","sources":["../../../src/rules/tech/robots-sitemap-presence.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,UAAU,EAAE,UAAU,EAAE,MAAM,gBAAgB,CAAC;AAc7D,wBAAsB,yBAAyB,CAAC,MAAM,EAAE,MAAM,GAAG,OAAO,CAAC,UAAU,EAAE,CAAC,CAqDrF;AAED;;;GAGG;AACH,6FAA6F;AAC7F,wBAAgB,sBAAsB,CAAC,SAAS,EAAE,MAAM,GAAG,MAAM,CAmBhE;AAED,wBAAgB,qBAAqB,CAAC,SAAS,EAAE,MAAM,GAAG,MAAM,EAAE,CA8BjE;AAED;;;;;;;GAOG;AACH,wBAAgB,kBAAkB,CAAC,OAAO,EAAE,MAAM,EAAE,OAAO,EAAE,MAAM,GAAG,OAAO,CA8B5E;AAED,wBAAgB,oBAAoB,CAClC,KAAK,EAAE,UAAU,EAAE,EACnB,WAAW,EAAE,GAAG,CAAC,MAAM,CAAC,EACxB,gBAAgB,EAAE,MAAM,GACvB,UAAU,EAAE,CA+Bd"}
+{"version":3,"file":"robots-sitemap-presence.d.ts","sourceRoot":"","sources":["../../../src/rules/tech/robots-sitemap-presence.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,UAAU,EAAE,UAAU,EAAE,MAAM,gBAAgB,CAAC;AAc7D,wBAAsB,yBAAyB,CAAC,MAAM,EAAE,MAAM,GAAG,OAAO,CAAC,UAAU,EAAE,CAAC,CAqDrF;AAED;;;GAGG;AACH,6FAA6F;AAC7F,wBAAgB,sBAAsB,CAAC,SAAS,EAAE,MAAM,GAAG,MAAM,CAmBhE;AAED;;;;;;;GAOG;AACH,wBAAgB,qBAAqB,CAAC,SAAS,EAAE,MAAM,EAAE,UAAU,GAAE,SAAS,MAAM,EAAU,GAAG,MAAM,EAAE,CA+BxG;AAED;;;;;;;GAOG;AACH,wBAAgB,kBAAkB,CAAC,OAAO,EAAE,MAAM,EAAE,OAAO,EAAE,MAAM,GAAG,OAAO,CA8B5E;AAED,wBAAgB,oBAAoB,CAClC,KAAK,EAAE,UAAU,EAAE,EACnB,WAAW,EAAE,GAAG,CAAC,MAAM,CAAC,EACxB,gBAAgB,EAAE,MAAM,GACvB,UAAU,EAAE,CA+Bd"}

package/dist/rules/tech/robots-sitemap-presence.js

@@ -86,10 +86,19 @@ export function parseCrawlDelaySeconds(robotsTxt) {
     }
     return 0;
 }
-export function parseDisallowPatterns(robotsTxt) {
+/**
+ * Parse Disallow patterns for the given user-agents. Merges the `User-agent: *`
+ * block with any named UA blocks; a hostile site that specifically disallows
+ * `pseolint` without a wildcard still gets honored.
+ *
+ * The default UA list (`["*"]`) preserves the pre-v0.3.2 behavior for callers
+ * that don't opt into UA-specific matching.
+ */
+export function parseDisallowPatterns(robotsTxt, userAgents = ["*"]) {
     const lines = robotsTxt.split(/\r?\n/);
     const patterns = [];
-    let inWildcardBlock = false;
+    const uaSet = new Set(userAgents.map((u) => u.toLowerCase()));
+    let inMatchingBlock = false;
     for (const rawLine of lines) {
         const line = rawLine.trim();
         // Skip comments and blank lines
@@ -97,11 +106,11 @@ export function parseDisallowPatterns(robotsTxt) {
             continue;
         // Detect User-agent directive
         if (/^user-agent\s*:/i.test(line)) {
-            const value = line.replace(/^user-agent\s*:\s*/i, "").trim();
-            inWildcardBlock = value === "*";
+            const value = line.replace(/^user-agent\s*:\s*/i, "").trim().toLowerCase();
+            inMatchingBlock = uaSet.has(value);
             continue;
         }
-        if (!inWildcardBlock)
+        if (!inMatchingBlock)
             continue;
         if (/^disallow\s*:/i.test(line)) {
             const value = line.replace(/^disallow\s*:\s*/i, "").trim();

package/dist/rules/tech/robots-sitemap-presence.js.map

@@ -1 +1 @@
-{"version":3,"file":"robots-sitemap-presence.js","sourceRoot":"","sources":["../../../src/rules/tech/robots-sitemap-presence.ts"],"names":[],"mappings":"AAEA,KAAK,UAAU,aAAa,CAAC,GAAW;IACtC,IAAI,CAAC;QACH,MAAM,QAAQ,GAAG,MAAM,KAAK,CAAC,GAAG,CAAC,CAAC;QAClC,IAAI,CAAC,QAAQ,CAAC,EAAE,EAAE,CAAC;YACjB,OAAO,IAAI,CAAC;QACd,CAAC;QACD,OAAO,MAAM,QAAQ,CAAC,IAAI,EAAE,CAAC;IAC/B,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,IAAI,CAAC;IACd,CAAC;AACH,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,yBAAyB,CAAC,MAAc;IAC5D,IAAI,CAAC,eAAe,CAAC,IAAI,CAAC,MAAM,CAAC,EAAE,CAAC;QAClC,OAAO,EAAE,CAAC;IACZ,CAAC;IAED,MAAM,SAAS,GAAG,IAAI,GAAG,CAAC,MAAM,CAAC,CAAC;IAClC,MAAM,MAAM,GAAG,SAAS,CAAC,MAAM,CAAC;IAChC,MAAM,SAAS,GAAG,GAAG,MAAM,aAAa,CAAC;IACzC,MAAM,UAAU,GAAG,GAAG,MAAM,cAAc,CAAC;IAC3C,MAAM,QAAQ,GAAiB,EAAE,CAAC;IAElC,MAAM,UAAU,GAAG,MAAM,aAAa,CAAC,SAAS,CAAC,CAAC;IAClD,IAAI,CAAC,UAAU,EAAE,CAAC;QAChB,QAAQ,CAAC,IAAI,CAAC;YACZ,MAAM,EAAE,8BAA8B;YACtC,QAAQ,EAAE,SAAS;YACnB,OAAO,EAAE,mBAAmB,SAAS,GAAG;YACxC,GAAG,EAAE,+BAA+B,SAAS,4DAA4D;SAC1G,CAAC,CAAC;QACH,OAAO,QAAQ,CAAC;IAClB,CAAC;IAED,IAAI,CAAC,oBAAoB,CAAC,IAAI,CAAC,UAAU,CAAC,EAAE,CAAC;QAC3C,QAAQ,CAAC,IAAI,CAAC;YACZ,MAAM,EAAE,8BAA8B;YACtC,QAAQ,EAAE,MAAM;YAChB,OAAO,EAAE,GAAG,SAAS,wCAAwC;YAC7D,GAAG,EAAE,8BAA8B,SAAS,oBAAoB,UAAU,EAAE;SAC7E,CAAC,CAAC;IACL,CAAC;IAED,MAAM,WAAW,GAAG,MAAM,aAAa,CAAC,UAAU,CAAC,CAAC;IACpD,IAAI,CAAC,WAAW,EAAE,CAAC;QACjB,QAAQ,CAAC,IAAI,CAAC;YACZ,MAAM,EAAE,8BAA8B;YACtC,QAAQ,EAAE,SAAS;YACnB,OAAO,EAAE,mBAAmB,UAAU,GAAG;YACzC,GAAG,EAAE,4BAA4B,UAAU,wBAAwB,SAAS,GAAG;SAChF,CAAC,CAAC;QACH,OAAO,QAAQ,CAAC;IAClB,CAAC;IAED,MAAM,OAAO,GAAG,WAAW,CAAC,WAAW,EAAE,CAAC;IAC1C,IAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,SAAS,CAAC,IAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,eAAe,CAAC,EAAE,CAAC;QACvE,QAAQ,CAAC,IAAI,CAAC;YACZ,MAAM,EAAE,8BAA8B;YACtC,QAAQ,EAAE,SAAS;YACnB,OAAO,EAAE,GAAG,UAAU,kDAAkD;YACxE,GAAG,EAAE,UAAU,UAAU,6EAA6E;SACvG,CAAC,CAAC;IACL,CAAC;IAED,OAAO,QAAQ,CAAC;AAClB,CAAC;AAED;;;GAGG;AACH,6FAA6F;AAC7F,MAAM,UAAU,sBAAsB,CAAC,SAAiB;IACtD,MAAM,KAAK,GAAG,SAAS,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;IACvC,IAAI,eAAe,GAAG,KAAK,CAAC;IAC5B,KAAK,MAAM,OAAO,IAAI,KAAK,EAAE,CAAC;QAC5B,MAAM,IAAI,GAAG,OAAO,CAAC,IAAI,EAAE,CAAC;QAC5B,IAAI,CAAC,IAAI,IAAI,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC;YAAE,SAAS;QAC5C,IAAI,kBAAkB,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;YAClC,MAAM,KAAK,GAAG,IAAI,CAAC,OAAO,CAAC,qBAAqB,EAAE,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC;YAC7D,eAAe,GAAG,KAAK,KAAK,GAAG,CAAC;YAChC,SAAS;QACX,CAAC;QACD,IAAI,CAAC,eAAe;YAAE,SAAS;QAC/B,IAAI,mBAAmB,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;YACnC,MAAM,KAAK,GAAG,IAAI,CAAC,OAAO,CAAC,sBAAsB,EAAE,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC;YAC9D,MAAM,CAAC,GAAG,MAAM,CAAC,KAAK,CAAC,CAAC;YACxB,IAAI,MAAM,CAAC,QAAQ,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC;gBAAE,OAAO,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;QAC1D,CAAC;IACH,CAAC;IACD,OAAO,CAAC,CAAC;AACX,CAAC;AAED,MAAM,UAAU,qBAAqB,CAAC,SAAiB;IACrD,MAAM,KAAK,GAAG,SAAS,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;IACvC,MAAM,QAAQ,GAAa,EAAE,CAAC;IAC9B,IAAI,eAAe,GAAG,KAAK,CAAC;IAE5B,KAAK,MAAM,OAAO,IAAI,KAAK,EAAE,CAAC;QAC5B,MAAM,IAAI,GAAG,OAAO,CAAC,IAAI,EAAE,CAAC;QAE5B,gCAAgC;QAChC,IAAI,CAAC,IAAI,IAAI,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC;YAAE,SAAS;QAE5C,8BAA8B;QAC9B,IAAI,kBAAkB,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;YAClC,MAAM,KAAK,GAAG,IAAI,CAAC,OAAO,CAAC,qBAAqB,EAAE,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC;YAC7D,eAAe,GAAG,KAAK,KAAK,GAAG,CAAC;YAChC,SAAS;QACX,CAAC;QAED,IAAI,CAAC,eAAe;YAAE,SAAS;QAE/B,IAAI,gBAAgB,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;YAChC,MAAM,KAAK,GAAG,IAAI,CAAC,OAAO,CAAC,mBAAmB,EAAE,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC;YAC3D,iDAAiD;YACjD,IAAI,KAAK,EAAE,CAAC;gBACV,QAAQ,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;YACvB,CAAC;QACH,CAAC;IACH,CAAC;IAED,OAAO,QAAQ,CAAC;AAClB,CAAC;AAED;;;;;;;GAOG;AACH,MAAM,UAAU,kBAAkB,CAAC,OAAe,EAAE,OAAe;IACjE,IAAI,CAAC,OAAO;QAAE,OAAO,KAAK,CAAC;IAE3B,IAAI,OAAO,CAAC,QAAQ,CAAC,GAAG,CAAC,EAAE,CAAC;QAC1B,yEAAyE;QACzE,2CAA2C;QAC3C,MAAM,KAAK,GAAG,OAAO,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;QACjC,IAAI,GAAG,GAAG,CAAC,CAAC;QACZ,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC;YACzC,MAAM,IAAI,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;YACtB,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC;gBACZ,8EAA8E;gBAC9E,IAAI,CAAC,OAAO,CAAC,UAAU,CAAC,IAAI,CAAC;oBAAE,OAAO,KAAK,CAAC;gBAC5C,GAAG,GAAG,IAAI,CAAC,MAAM,CAAC;YACpB,CAAC;iBAAM,CAAC;gBACN,MAAM,GAAG,GAAG,OAAO,CAAC,OAAO,CAAC,IAAI,EAAE,GAAG,CAAC,CAAC;gBACvC,IAAI,GAAG,KAAK,CAAC,CAAC;oBAAE,OAAO,KAAK,CAAC;gBAC7B,GAAG,GAAG,GAAG,GAAG,IAAI,CAAC,MAAM,CAAC;YAC1B,CAAC;QACH,CAAC;QACD,OAAO,IAAI,CAAC;IACd,CAAC;IAED,+BAA+B;IAC/B,IAAI,OAAO,CAAC,QAAQ,CAAC,GAAG,CAAC,EAAE,CAAC;QAC1B,OAAO,OAAO,CAAC,UAAU,CAAC,OAAO,CAAC,CAAC;IACrC,CAAC;IAED,cAAc;IACd,OAAO,OAAO,KAAK,OAAO,CAAC;AAC7B,CAAC;AAED,MAAM,UAAU,oBAAoB,CAClC,KAAmB,EACnB,WAAwB,EACxB,gBAAwB;IAExB,MAAM,gBAAgB,GAAG,qBAAqB,CAAC,gBAAgB,CAAC,CAAC;IACjE,IAAI,gBAAgB,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,EAAE,CAAC;IAE7C,MAAM,QAAQ,GAAiB,EAAE,CAAC;IAElC,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,IAAI,CAAC,WAAW,CAAC,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC;YAAE,SAAS;QAEzC,IAAI,UAAkB,CAAC;QACvB,IAAI,CAAC;YACH,UAAU,GAAG,IAAI,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,QAAQ,CAAC;QAC1C,CAAC;QAAC,MAAM,CAAC;YACP,SAAS;QACX,CAAC;QAED,KAAK,MAAM,OAAO,IAAI,gBAAgB,EAAE,CAAC;YACvC,IAAI,kBAAkB,CAAC,UAAU,EAAE,OAAO,CAAC,EAAE,CAAC;gBAC5C,QAAQ,CAAC,IAAI,CAAC;oBACZ,MAAM,EAAE,wBAAwB;oBAChC,QAAQ,EAAE,OAAO;oBACjB,OAAO,EAAE,GAAG,IAAI,CAAC,GAAG,2DAA2D,OAAO,IAAI;oBAC1F,OAAO,EAAE,IAAI,CAAC,GAAG;oBACjB,GAAG,EAAE,qEAAqE;iBAC3E,CAAC,CAAC;gBACH,MAAM,CAAC,iCAAiC;YAC1C,CAAC;QACH,CAAC;IACH,CAAC;IAED,OAAO,QAAQ,CAAC;AAClB,CAAC"}
+{"version":3,"file":"robots-sitemap-presence.js","sourceRoot":"","sources":["../../../src/rules/tech/robots-sitemap-presence.ts"],"names":[],"mappings":"AAEA,KAAK,UAAU,aAAa,CAAC,GAAW;IACtC,IAAI,CAAC;QACH,MAAM,QAAQ,GAAG,MAAM,KAAK,CAAC,GAAG,CAAC,CAAC;QAClC,IAAI,CAAC,QAAQ,CAAC,EAAE,EAAE,CAAC;YACjB,OAAO,IAAI,CAAC;QACd,CAAC;QACD,OAAO,MAAM,QAAQ,CAAC,IAAI,EAAE,CAAC;IAC/B,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,IAAI,CAAC;IACd,CAAC;AACH,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,yBAAyB,CAAC,MAAc;IAC5D,IAAI,CAAC,eAAe,CAAC,IAAI,CAAC,MAAM,CAAC,EAAE,CAAC;QAClC,OAAO,EAAE,CAAC;IACZ,CAAC;IAED,MAAM,SAAS,GAAG,IAAI,GAAG,CAAC,MAAM,CAAC,CAAC;IAClC,MAAM,MAAM,GAAG,SAAS,CAAC,MAAM,CAAC;IAChC,MAAM,SAAS,GAAG,GAAG,MAAM,aAAa,CAAC;IACzC,MAAM,UAAU,GAAG,GAAG,MAAM,cAAc,CAAC;IAC3C,MAAM,QAAQ,GAAiB,EAAE,CAAC;IAElC,MAAM,UAAU,GAAG,MAAM,aAAa,CAAC,SAAS,CAAC,CAAC;IAClD,IAAI,CAAC,UAAU,EAAE,CAAC;QAChB,QAAQ,CAAC,IAAI,CAAC;YACZ,MAAM,EAAE,8BAA8B;YACtC,QAAQ,EAAE,SAAS;YACnB,OAAO,EAAE,mBAAmB,SAAS,GAAG;YACxC,GAAG,EAAE,+BAA+B,SAAS,4DAA4D;SAC1G,CAAC,CAAC;QACH,OAAO,QAAQ,CAAC;IAClB,CAAC;IAED,IAAI,CAAC,oBAAoB,CAAC,IAAI,CAAC,UAAU,CAAC,EAAE,CAAC;QAC3C,QAAQ,CAAC,IAAI,CAAC;YACZ,MAAM,EAAE,8BAA8B;YACtC,QAAQ,EAAE,MAAM;YAChB,OAAO,EAAE,GAAG,SAAS,wCAAwC;YAC7D,GAAG,EAAE,8BAA8B,SAAS,oBAAoB,UAAU,EAAE;SAC7E,CAAC,CAAC;IACL,CAAC;IAED,MAAM,WAAW,GAAG,MAAM,aAAa,CAAC,UAAU,CAAC,CAAC;IACpD,IAAI,CAAC,WAAW,EAAE,CAAC;QACjB,QAAQ,CAAC,IAAI,CAAC;YACZ,MAAM,EAAE,8BAA8B;YACtC,QAAQ,EAAE,SAAS;YACnB,OAAO,EAAE,mBAAmB,UAAU,GAAG;YACzC,GAAG,EAAE,4BAA4B,UAAU,wBAAwB,SAAS,GAAG;SAChF,CAAC,CAAC;QACH,OAAO,QAAQ,CAAC;IAClB,CAAC;IAED,MAAM,OAAO,GAAG,WAAW,CAAC,WAAW,EAAE,CAAC;IAC1C,IAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,SAAS,CAAC,IAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,eAAe,CAAC,EAAE,CAAC;QACvE,QAAQ,CAAC,IAAI,CAAC;YACZ,MAAM,EAAE,8BAA8B;YACtC,QAAQ,EAAE,SAAS;YACnB,OAAO,EAAE,GAAG,UAAU,kDAAkD;YACxE,GAAG,EAAE,UAAU,UAAU,6EAA6E;SACvG,CAAC,CAAC;IACL,CAAC;IAED,OAAO,QAAQ,CAAC;AAClB,CAAC;AAED;;;GAGG;AACH,6FAA6F;AAC7F,MAAM,UAAU,sBAAsB,CAAC,SAAiB;IACtD,MAAM,KAAK,GAAG,SAAS,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;IACvC,IAAI,eAAe,GAAG,KAAK,CAAC;IAC5B,KAAK,MAAM,OAAO,IAAI,KAAK,EAAE,CAAC;QAC5B,MAAM,IAAI,GAAG,OAAO,CAAC,IAAI,EAAE,CAAC;QAC5B,IAAI,CAAC,IAAI,IAAI,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC;YAAE,SAAS;QAC5C,IAAI,kBAAkB,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;YAClC,MAAM,KAAK,GAAG,IAAI,CAAC,OAAO,CAAC,qBAAqB,EAAE,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC;YAC7D,eAAe,GAAG,KAAK,KAAK,GAAG,CAAC;YAChC,SAAS;QACX,CAAC;QACD,IAAI,CAAC,eAAe;YAAE,SAAS;QAC/B,IAAI,mBAAmB,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;YACnC,MAAM,KAAK,GAAG,IAAI,CAAC,OAAO,CAAC,sBAAsB,EAAE,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC;YAC9D,MAAM,CAAC,GAAG,MAAM,CAAC,KAAK,CAAC,CAAC;YACxB,IAAI,MAAM,CAAC,QAAQ,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC;gBAAE,OAAO,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;QAC1D,CAAC;IACH,CAAC;IACD,OAAO,CAAC,CAAC;AACX,CAAC;AAED;;;;;;;GAOG;AACH,MAAM,UAAU,qBAAqB,CAAC,SAAiB,EAAE,aAAgC,CAAC,GAAG,CAAC;IAC5F,MAAM,KAAK,GAAG,SAAS,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;IACvC,MAAM,QAAQ,GAAa,EAAE,CAAC;IAC9B,MAAM,KAAK,GAAG,IAAI,GAAG,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,WAAW,EAAE,CAAC,CAAC,CAAC;IAC9D,IAAI,eAAe,GAAG,KAAK,CAAC;IAE5B,KAAK,MAAM,OAAO,IAAI,KAAK,EAAE,CAAC;QAC5B,MAAM,IAAI,GAAG,OAAO,CAAC,IAAI,EAAE,CAAC;QAE5B,gCAAgC;QAChC,IAAI,CAAC,IAAI,IAAI,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC;YAAE,SAAS;QAE5C,8BAA8B;QAC9B,IAAI,kBAAkB,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;YAClC,MAAM,KAAK,GAAG,IAAI,CAAC,OAAO,CAAC,qBAAqB,EAAE,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC;YAC3E,eAAe,GAAG,KAAK,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC;YACnC,SAAS;QACX,CAAC;QAED,IAAI,CAAC,eAAe;YAAE,SAAS;QAE/B,IAAI,gBAAgB,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;YAChC,MAAM,KAAK,GAAG,IAAI,CAAC,OAAO,CAAC,mBAAmB,EAAE,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC;YAC3D,iDAAiD;YACjD,IAAI,KAAK,EAAE,CAAC;gBACV,QAAQ,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;YACvB,CAAC;QACH,CAAC;IACH,CAAC;IAED,OAAO,QAAQ,CAAC;AAClB,CAAC;AAED;;;;;;;GAOG;AACH,MAAM,UAAU,kBAAkB,CAAC,OAAe,EAAE,OAAe;IACjE,IAAI,CAAC,OAAO;QAAE,OAAO,KAAK,CAAC;IAE3B,IAAI,OAAO,CAAC,QAAQ,CAAC,GAAG,CAAC,EAAE,CAAC;QAC1B,yEAAyE;QACzE,2CAA2C;QAC3C,MAAM,KAAK,GAAG,OAAO,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;QACjC,IAAI,GAAG,GAAG,CAAC,CAAC;QACZ,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC;YACzC,MAAM,IAAI,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;YACtB,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC;gBACZ,8EAA8E;gBAC9E,IAAI,CAAC,OAAO,CAAC,UAAU,CAAC,IAAI,CAAC;oBAAE,OAAO,KAAK,CAAC;gBAC5C,GAAG,GAAG,IAAI,CAAC,MAAM,CAAC;YACpB,CAAC;iBAAM,CAAC;gBACN,MAAM,GAAG,GAAG,OAAO,CAAC,OAAO,CAAC,IAAI,EAAE,GAAG,CAAC,CAAC;gBACvC,IAAI,GAAG,KAAK,CAAC,CAAC;oBAAE,OAAO,KAAK,CAAC;gBAC7B,GAAG,GAAG,GAAG,GAAG,IAAI,CAAC,MAAM,CAAC;YAC1B,CAAC;QACH,CAAC;QACD,OAAO,IAAI,CAAC;IACd,CAAC;IAED,+BAA+B;IAC/B,IAAI,OAAO,CAAC,QAAQ,CAAC,GAAG,CAAC,EAAE,CAAC;QAC1B,OAAO,OAAO,CAAC,UAAU,CAAC,OAAO,CAAC,CAAC;IACrC,CAAC;IAED,cAAc;IACd,OAAO,OAAO,KAAK,OAAO,CAAC;AAC7B,CAAC;AAED,MAAM,UAAU,oBAAoB,CAClC,KAAmB,EACnB,WAAwB,EACxB,gBAAwB;IAExB,MAAM,gBAAgB,GAAG,qBAAqB,CAAC,gBAAgB,CAAC,CAAC;IACjE,IAAI,gBAAgB,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,EAAE,CAAC;IAE7C,MAAM,QAAQ,GAAiB,EAAE,CAAC;IAElC,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,IAAI,CAAC,WAAW,CAAC,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC;YAAE,SAAS;QAEzC,IAAI,UAAkB,CAAC;QACvB,IAAI,CAAC;YACH,UAAU,GAAG,IAAI,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,QAAQ,CAAC;QAC1C,CAAC;QAAC,MAAM,CAAC;YACP,SAAS;QACX,CAAC;QAED,KAAK,MAAM,OAAO,IAAI,gBAAgB,EAAE,CAAC;YACvC,IAAI,kBAAkB,CAAC,UAAU,EAAE,OAAO,CAAC,EAAE,CAAC;gBAC5C,QAAQ,CAAC,IAAI,CAAC;oBACZ,MAAM,EAAE,wBAAwB;oBAChC,QAAQ,EAAE,OAAO;oBACjB,OAAO,EAAE,GAAG,IAAI,CAAC,GAAG,2DAA2D,OAAO,IAAI;oBAC1F,OAAO,EAAE,IAAI,CAAC,GAAG;oBACjB,GAAG,EAAE,qEAAqE;iBAC3E,CAAC,CAAC;gBACH,MAAM,CAAC,iCAAiC;YAC1C,CAAC;QACH,CAAC;IACH,CAAC;IAED,OAAO,QAAQ,CAAC;AAClB,CAAC"}

package/dist/safe-mode-preset.d.ts

@@ -0,0 +1,27 @@
+import type { SafeMode, AuditOptions } from "./types.js";
+export type SafeModePreset = {
+    concurrency?: number;
+    sampleSize?: number;
+    guardSsrf?: boolean;
+    respectRobotsTxt?: boolean;
+    followRedirects?: boolean;
+    maxCrawlDiscovered?: number;
+    maxFetchBytes?: number;
+};
+export type SafeModeKey = SafeMode | "__none";
+/**
+ * Presets that flip several safety defaults at once. Individual options on
+ * AuditOptions override the preset when explicitly set. `__none` is the
+ * sentinel key used when no preset is selected — all fields undefined so the
+ * `??` chain falls through to hard-coded defaults.
+ */
+export declare const SAFE_MODE_PRESETS: Record<SafeModeKey, SafeModePreset>;
+/**
+ * Pick the preset key for a given source + options combination.
+ *
+ *   explicit `safeMode`  → that preset
+ *   autoDevPreset !== false && localhost URL → "dev"
+ *   otherwise            → "__none"
+ */
+export declare function resolveSafeModeKey(source: string, options?: Pick<AuditOptions, "safeMode" | "autoDevPreset">): SafeModeKey;
+//# sourceMappingURL=safe-mode-preset.d.ts.map

package/dist/safe-mode-preset.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"safe-mode-preset.d.ts","sourceRoot":"","sources":["../src/safe-mode-preset.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,QAAQ,EAAE,YAAY,EAAE,MAAM,YAAY,CAAC;AAGzD,MAAM,MAAM,cAAc,GAAG;IAC3B,WAAW,CAAC,EAAE,MAAM,CAAC;IACrB,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,SAAS,CAAC,EAAE,OAAO,CAAC;IACpB,gBAAgB,CAAC,EAAE,OAAO,CAAC;IAC3B,eAAe,CAAC,EAAE,OAAO,CAAC;IAC1B,kBAAkB,CAAC,EAAE,MAAM,CAAC;IAC5B,aAAa,CAAC,EAAE,MAAM,CAAC;CACxB,CAAC;AAEF,MAAM,MAAM,WAAW,GAAG,QAAQ,GAAG,QAAQ,CAAC;AAE9C;;;;;GAKG;AACH,eAAO,MAAM,iBAAiB,EAAE,MAAM,CAAC,WAAW,EAAE,cAAc,CA6BjE,CAAC;AAEF;;;;;;GAMG;AACH,wBAAgB,kBAAkB,CAChC,MAAM,EAAE,MAAM,EACd,OAAO,CAAC,EAAE,IAAI,CAAC,YAAY,EAAE,UAAU,GAAG,eAAe,CAAC,GACzD,WAAW,CAKb"}

package/dist/safe-mode-preset.js

@@ -0,0 +1,54 @@
+import { isLocalhostUrl } from "./ssrf-guard.js";
+/**
+ * Presets that flip several safety defaults at once. Individual options on
+ * AuditOptions override the preset when explicitly set. `__none` is the
+ * sentinel key used when no preset is selected — all fields undefined so the
+ * `??` chain falls through to hard-coded defaults.
+ */
+export const SAFE_MODE_PRESETS = {
+    saas: {
+        // Hosted-service defaults. Assume user-submitted URLs are hostile.
+        guardSsrf: true,
+        respectRobotsTxt: true,
+        followRedirects: true,
+        maxCrawlDiscovered: 2000,
+        maxFetchBytes: 10_000_000,
+    },
+    cli: {
+        // Local / dev defaults. User auditing their own or local site.
+        guardSsrf: false,
+        respectRobotsTxt: true,
+        followRedirects: true,
+        maxCrawlDiscovered: 5000,
+        maxFetchBytes: 52_428_800,
+    },
+    dev: {
+        // Localhost probe. A cache-cold dev server can amplify every crawled URL
+        // into a chain of DB queries — keep the first pass tiny and reversible.
+        concurrency: 1,
+        sampleSize: 25,
+        guardSsrf: false,
+        respectRobotsTxt: true,
+        followRedirects: true,
+        maxCrawlDiscovered: 50,
+        maxFetchBytes: 52_428_800,
+    },
+    __none: {},
+};
+/**
+ * Pick the preset key for a given source + options combination.
+ *
+ *   explicit `safeMode`  → that preset
+ *   autoDevPreset !== false && localhost URL → "dev"
+ *   otherwise            → "__none"
+ */
+export function resolveSafeModeKey(source, options) {
+    if (options?.safeMode)
+        return options.safeMode;
+    if (options?.autoDevPreset === false)
+        return "__none";
+    if (isLocalhostUrl(source))
+        return "dev";
+    return "__none";
+}
+//# sourceMappingURL=safe-mode-preset.js.map

package/dist/safe-mode-preset.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"safe-mode-preset.js","sourceRoot":"","sources":["../src/safe-mode-preset.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,cAAc,EAAE,MAAM,iBAAiB,CAAC;AAcjD;;;;;GAKG;AACH,MAAM,CAAC,MAAM,iBAAiB,GAAwC;IACpE,IAAI,EAAE;QACJ,mEAAmE;QACnE,SAAS,EAAE,IAAI;QACf,gBAAgB,EAAE,IAAI;QACtB,eAAe,EAAE,IAAI;QACrB,kBAAkB,EAAE,IAAI;QACxB,aAAa,EAAE,UAAU;KAC1B;IACD,GAAG,EAAE;QACH,+DAA+D;QAC/D,SAAS,EAAE,KAAK;QAChB,gBAAgB,EAAE,IAAI;QACtB,eAAe,EAAE,IAAI;QACrB,kBAAkB,EAAE,IAAI;QACxB,aAAa,EAAE,UAAU;KAC1B;IACD,GAAG,EAAE;QACH,yEAAyE;QACzE,wEAAwE;QACxE,WAAW,EAAE,CAAC;QACd,UAAU,EAAE,EAAE;QACd,SAAS,EAAE,KAAK;QAChB,gBAAgB,EAAE,IAAI;QACtB,eAAe,EAAE,IAAI;QACrB,kBAAkB,EAAE,EAAE;QACtB,aAAa,EAAE,UAAU;KAC1B;IACD,MAAM,EAAE,EAAE;CACX,CAAC;AAEF;;;;;;GAMG;AACH,MAAM,UAAU,kBAAkB,CAChC,MAAc,EACd,OAA0D;IAE1D,IAAI,OAAO,EAAE,QAAQ;QAAE,OAAO,OAAO,CAAC,QAAQ,CAAC;IAC/C,IAAI,OAAO,EAAE,aAAa,KAAK,KAAK;QAAE,OAAO,QAAQ,CAAC;IACtD,IAAI,cAAc,CAAC,MAAM,CAAC;QAAE,OAAO,KAAK,CAAC;IACzC,OAAO,QAAQ,CAAC;AAClB,CAAC"}

package/dist/site-classifier.d.ts

@@ -0,0 +1,83 @@
+/**
+ * v0.4 Site Classifier — runs BEFORE rules to gate which rule set applies.
+ *
+ * Today the engine runs every rule against every audited site regardless of
+ * whether the site is actually programmatic-SEO. A 23-page marketing site
+ * gets pSEO-targeted findings that don't apply; a 50,000-page directory
+ * gets the same audit shape as a small blog. This pre-flight classifier
+ * decides which "kind" of site we're auditing and emits a list of ruleIds
+ * that the dispatcher will skip when site type is small-marketing/blog.
+ *
+ * Heuristics (v1, ships in v0.4):
+ *   - Sitemap URL count: <50 = bias small/blog; ≥1000 = bias programmatic.
+ *   - URL-pattern clustering: normalize numeric/slug segments; if top-3
+ *     templates cover ≥60% of URLs → strong programmatic signal.
+ *   - Framework signal: pass-through from dev-server detection.
+ *
+ * Deferred to v0.4.1+ (per spec §4.11):
+ *   - DOM-skeleton hashing across pages.
+ *   - Per-cluster classification (mixed pSEO + marketing).
+ *   - Per-page applicability tagging on findings.
+ */
+export type SiteType = "programmatic-directory" | "small-marketing" | "blog" | "ecommerce" | "unclear";
+export type ClassificationSignal = {
+    kind: "sitemap-url-count";
+    value: number;
+} | {
+    kind: "url-pattern-cluster-coverage";
+    topTemplate: string;
+    pages: number;
+    ratio: number;
+} | {
+    kind: "framework-detected";
+    value: "nextjs" | "vite" | "astro" | "unknown";
+};
+export interface SiteClassification {
+    type: SiteType;
+    /** 0–1 confidence. Below 0.6 → treat as `unclear` and run all rules. */
+    confidence: number;
+    /** Ordered observations that fed the classification. UI surfaces these. */
+    signals: ClassificationSignal[];
+    /**
+     * RuleIds suppressed because of this classification. The dispatcher
+     * checks this list before invoking each rule.
+     *
+     * Empty array when type is `programmatic-directory`, `ecommerce`, or
+     * `unclear` — those run all rules.
+     */
+    suppressedRules: string[];
+}
+/** Rules suppressed for non-pSEO sites (small-marketing / blog). */
+export declare const PSEO_ONLY_RULE_IDS: readonly string[];
+/**
+ * Normalize a pathname into a hashable template by replacing path segments
+ * that look like values with type placeholders.
+ *
+ * Examples:
+ *   /california/los-angeles/plumbers   → /:slug/:slug/:slug
+ *   /blog/hello-world                  → /blog/:slug
+ *   /post/12345                        → /post/:n
+ *   /                                  → /
+ */
+export declare function normalizePathToTemplate(pathname: string): string;
+/** Compute template-cluster ratios from a URL list. Returns top entries first. */
+export declare function clusterUrlTemplates(urls: string[]): Array<{
+    template: string;
+    count: number;
+    ratio: number;
+}>;
+export interface ClassifySiteInput {
+    /** All discovered URLs (sitemap + crawl). Unfiltered. */
+    urls: string[];
+    /** Framework detected via dev-server response headers, if any. */
+    framework?: "nextjs" | "vite" | "astro" | "unknown";
+}
+/**
+ * Classify a site from its URL list + framework signal. Pure function.
+ *
+ * Contract: callers must pass the FULL discovered URL list (sitemap +
+ * crawl), not the post-sample list. The classifier needs the raw size
+ * signal to distinguish a 5000-page directory from a 25-page sample of one.
+ */
+export declare function classifySite(input: ClassifySiteInput): SiteClassification;
+//# sourceMappingURL=site-classifier.d.ts.map

package/dist/site-classifier.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"site-classifier.d.ts","sourceRoot":"","sources":["../src/site-classifier.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;GAoBG;AAEH,MAAM,MAAM,QAAQ,GAChB,wBAAwB,GACxB,iBAAiB,GACjB,MAAM,GACN,WAAW,GACX,SAAS,CAAC;AAEd,MAAM,MAAM,oBAAoB,GAC5B;IAAE,IAAI,EAAE,mBAAmB,CAAC;IAAC,KAAK,EAAE,MAAM,CAAA;CAAE,GAC5C;IACE,IAAI,EAAE,8BAA8B,CAAC;IACrC,WAAW,EAAE,MAAM,CAAC;IACpB,KAAK,EAAE,MAAM,CAAC;IACd,KAAK,EAAE,MAAM,CAAC;CACf,GACD;IAAE,IAAI,EAAE,oBAAoB,CAAC;IAAC,KAAK,EAAE,QAAQ,GAAG,MAAM,GAAG,OAAO,GAAG,SAAS,CAAA;CAAE,CAAC;AAEnF,MAAM,WAAW,kBAAkB;IACjC,IAAI,EAAE,QAAQ,CAAC;IACf,wEAAwE;IACxE,UAAU,EAAE,MAAM,CAAC;IACnB,2EAA2E;IAC3E,OAAO,EAAE,oBAAoB,EAAE,CAAC;IAChC;;;;;;OAMG;IACH,eAAe,EAAE,MAAM,EAAE,CAAC;CAC3B;AAED,oEAAoE;AACpE,eAAO,MAAM,kBAAkB,EAAE,SAAS,MAAM,EAK/C,CAAC;AAEF;;;;;;;;;GASG;AACH,wBAAgB,uBAAuB,CAAC,QAAQ,EAAE,MAAM,GAAG,MAAM,CAiChE;AAiBD,kFAAkF;AAClF,wBAAgB,mBAAmB,CAAC,IAAI,EAAE,MAAM,EAAE,GAAG,KAAK,CAAC;IAAE,QAAQ,EAAE,MAAM,CAAC;IAAC,KAAK,EAAE,MAAM,CAAC;IAAC,KAAK,EAAE,MAAM,CAAA;CAAE,CAAC,CAiB7G;AAED,MAAM,WAAW,iBAAiB;IAChC,yDAAyD;IACzD,IAAI,EAAE,MAAM,EAAE,CAAC;IACf,kEAAkE;IAClE,SAAS,CAAC,EAAE,QAAQ,GAAG,MAAM,GAAG,OAAO,GAAG,SAAS,CAAC;CACrD;AAED;;;;;;GAMG;AACH,wBAAgB,YAAY,CAAC,KAAK,EAAE,iBAAiB,GAAG,kBAAkB,CA6FzE"}