@taodigital/crawler 0.0.0-stage → 1.2.0

package/dist/incremental-diff.js

@@ -0,0 +1,137 @@
+function hashContent(text) {
+  let h1 = 5381;
+  let h2 = 52711;
+  for (let i = 0; i < text.length; i++) {
+    const c = text.charCodeAt(i);
+    h1 = (Math.imul(h1, 33) ^ c) >>> 0;
+    h2 = (Math.imul(h2, 31) ^ c) >>> 0;
+  }
+  return (h1 >>> 0).toString(16).padStart(8, "0") + (h2 >>> 0).toString(16).padStart(8, "0");
+}
+function normalizeText(text) {
+  return text.split("\n").map((l) => l.trim()).filter((l) => l.length > 0).join("\n");
+}
+function computeContentHash(url, markdownText) {
+  const normalized = normalizeText(markdownText);
+  return {
+    url,
+    hash: hashContent(normalized),
+    textLength: normalized.length,
+    computedAt: (/* @__PURE__ */ new Date()).toISOString()
+  };
+}
+function diffContent(url, currentMarkdown, priorHashes) {
+  const current = computeContentHash(url, currentMarkdown);
+  const prior = priorHashes.get(url);
+  if (!prior) {
+    return { url, changed: true, currentHash: current.hash, reason: "new" };
+  }
+  if (prior === current.hash) {
+    return { url, changed: false, previousHash: prior, currentHash: current.hash, reason: "unchanged" };
+  }
+  return { url, changed: true, previousHash: prior, currentHash: current.hash, reason: "changed" };
+}
+function buildPriorHashMap(screenRecords) {
+  const map = /* @__PURE__ */ new Map();
+  for (const s of screenRecords) {
+    if (s.url && s.elementHash) map.set(s.url, s.elementHash);
+  }
+  return map;
+}
+function partitionByChange(urls, priorHashes) {
+  const toProcess = [];
+  const toSkip = [];
+  for (const url of urls) {
+    if (priorHashes.has(url)) {
+      toProcess.push(url);
+    } else {
+      toProcess.push(url);
+    }
+  }
+  return { toProcess, toSkip };
+}
+function computeTextDiffs(previous, current) {
+  const prevLines = previous.split("\n");
+  const currLines = current.split("\n");
+  const changes = [];
+  const prevSet = new Set(prevLines);
+  const currSet = new Set(currLines);
+  let lineNum = 0;
+  for (const line of prevLines) {
+    if (!currSet.has(line)) {
+      changes.push({ type: "removed", line: lineNum, content: line });
+    }
+    lineNum++;
+  }
+  lineNum = 0;
+  for (const line of currLines) {
+    if (!prevSet.has(line)) {
+      changes.push({ type: "added", line: lineNum, content: line });
+    }
+    lineNum++;
+  }
+  return changes;
+}
+function computeSetDiffs(previous, current) {
+  const prevSet = new Set(previous);
+  const currSet = new Set(current);
+  return {
+    added: current.filter((u) => !prevSet.has(u)),
+    removed: previous.filter((u) => !currSet.has(u))
+  };
+}
+function flattenJson(obj, prefix = "") {
+  if (typeof obj !== "object" || obj === null) {
+    return { [prefix]: obj };
+  }
+  const result = {};
+  for (const [k, v] of Object.entries(obj)) {
+    const path = prefix ? `${prefix}.${k}` : k;
+    if (typeof v === "object" && v !== null && !Array.isArray(v)) {
+      Object.assign(result, flattenJson(v, path));
+    } else {
+      result[path] = v;
+    }
+  }
+  return result;
+}
+function computeJsonDiffs(previous, current) {
+  const prevFlat = flattenJson(previous);
+  const currFlat = flattenJson(current);
+  const allKeys = /* @__PURE__ */ new Set([...Object.keys(prevFlat), ...Object.keys(currFlat)]);
+  const changes = [];
+  for (const path of allKeys) {
+    const oldVal = prevFlat[path];
+    const newVal = currFlat[path];
+    if (JSON.stringify(oldVal) !== JSON.stringify(newVal)) {
+      changes.push({ path, old: oldVal, new: newVal });
+    }
+  }
+  return changes;
+}
+function countMonitorDiffs(diffs) {
+  let count = 0;
+  if (diffs.markdown) count += diffs.markdown.length;
+  if (diffs.html) count += diffs.html.length;
+  if (diffs.json) count += diffs.json.length;
+  if (diffs.summary) count += diffs.summary.length;
+  if (diffs.links) count += diffs.links.added.length + diffs.links.removed.length;
+  if (diffs.images) count += diffs.images.added.length + diffs.images.removed.length;
+  if (diffs.branding) count += diffs.branding.length;
+  if (diffs.screenshot) count += diffs.screenshot.changed > 0 ? 1 : 0;
+  return count;
+}
+export {
+  buildPriorHashMap,
+  computeContentHash,
+  computeJsonDiffs,
+  computeSetDiffs,
+  computeTextDiffs,
+  countMonitorDiffs,
+  diffContent,
+  flattenJson,
+  hashContent,
+  normalizeText,
+  partitionByChange
+};
+//# sourceMappingURL=incremental-diff.js.map

package/dist/incremental-diff.js.map

@@ -0,0 +1,7 @@
+{
+  "version": 3,
+  "sources": ["../src/incremental-diff.ts"],
+  "sourcesContent": ["/**\n * Incremental crawl diff \u2014 compute stable content hashes and skip unchanged screens.\n *\n * Each crawled page gets a content fingerprint (murmurhash-free, pure JS).\n * On the next crawl of the same project, pages with the same fingerprint are\n * marked SKIPPED instead of re-processed, drastically reducing crawl time and\n * LLM token spend for sites with mostly-static content.\n *\n * Design: \"same content hash\" means the meaningful text hasn't changed.\n * Layout/CSS/JS changes that don't affect visible text are ignored intentionally.\n */\n\nexport interface ContentHash {\n  url: string;\n  hash: string;\n  textLength: number;\n  computedAt: string; // ISO timestamp\n}\n\nexport interface DiffResult {\n  url: string;\n  changed: boolean;\n  previousHash?: string;\n  currentHash: string;\n  reason: 'new' | 'changed' | 'unchanged';\n}\n\n/**\n * Compute a stable 64-bit-ish hash from text content.\n * djb2 variant \u2014 fast, no deps, collision-resistant enough for content fingerprinting.\n */\nexport function hashContent(text: string): string {\n  let h1 = 5381;\n  let h2 = 52711;\n  for (let i = 0; i < text.length; i++) {\n    const c = text.charCodeAt(i);\n    h1 = (Math.imul(h1, 33) ^ c) >>> 0;\n    h2 = (Math.imul(h2, 31) ^ c) >>> 0;\n  }\n  return (h1 >>> 0).toString(16).padStart(8, '0') + (h2 >>> 0).toString(16).padStart(8, '0');\n}\n\n/**\n * Normalize text before hashing \u2014 strips whitespace-only lines, trims each line,\n * collapses runs of blank lines. Makes hash stable across minor formatting changes.\n */\nexport function normalizeText(text: string): string {\n  return text\n    .split('\\n')\n    .map(l => l.trim())\n    .filter(l => l.length > 0)\n    .join('\\n');\n}\n\nexport function computeContentHash(url: string, markdownText: string): ContentHash {\n  const normalized = normalizeText(markdownText);\n  return {\n    url,\n    hash: hashContent(normalized),\n    textLength: normalized.length,\n    computedAt: new Date().toISOString(),\n  };\n}\n\n/**\n * Compare a page's current content against prior hashes.\n * Returns { changed: false } for unchanged pages \u2014 caller should skip re-processing.\n */\nexport function diffContent(\n  url: string,\n  currentMarkdown: string,\n  priorHashes: Map<string, string>, // url \u2192 hash\n): DiffResult {\n  const current = computeContentHash(url, currentMarkdown);\n  const prior = priorHashes.get(url);\n\n  if (!prior) {\n    return { url, changed: true, currentHash: current.hash, reason: 'new' };\n  }\n  if (prior === current.hash) {\n    return { url, changed: false, previousHash: prior, currentHash: current.hash, reason: 'unchanged' };\n  }\n  return { url, changed: true, previousHash: prior, currentHash: current.hash, reason: 'changed' };\n}\n\n/**\n * Build a hash map from prior screen records (from DB query result).\n * screenRecords: Array<{ url: string; elementHash?: string | null }>\n */\nexport function buildPriorHashMap(screenRecords: Array<{ url: string; elementHash?: string | null }>): Map<string, string> {\n  const map = new Map<string, string>();\n  for (const s of screenRecords) {\n    if (s.url && s.elementHash) map.set(s.url, s.elementHash);\n  }\n  return map;\n}\n\n/**\n * Filter a batch of discovered URLs against prior hashes.\n * Returns { toProcess, toSkip } \u2014 caller enqueues toProcess, logs toSkip count.\n *\n * Note: filtering at URL level only works if the prior crawl already visited the URL.\n * For new URLs, they're always in toProcess.\n */\nexport function partitionByChange(\n  urls: string[],\n  priorHashes: Map<string, string>,\n): { toProcess: string[]; toSkip: string[] } {\n  const toProcess: string[] = [];\n  const toSkip: string[] = [];\n  for (const url of urls) {\n    if (priorHashes.has(url)) {\n      // Can't know if changed without fetching \u2014 still needs fetch, but hash after fetch\n      // will determine skip. Include all \u2014 hash check happens post-fetch.\n      toProcess.push(url);\n    } else {\n      toProcess.push(url);\n    }\n  }\n  return { toProcess, toSkip };\n}\n\n// \u2500\u2500\u2500 Typed monitor diffs (GAP-01) \u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\n\n/** A single line-level change in text content (markdown, html, summary). */\nexport interface TextChange {\n  type: 'added' | 'removed';\n  line: number;\n  content: string;\n}\n\n/** A change to a specific JSON path between two scraped JSON objects. */\nexport interface JsonChange {\n  path: string;\n  old: unknown;\n  new: unknown;\n}\n\n/** Set-based change for collections like links or images. */\nexport interface SetChange {\n  added: string[];\n  removed: string[];\n}\n\n/** Screenshot pixel-level change. changed is percentage (0-100). */\nexport interface ImageChange {\n  size: number;\n  changed: number;\n  mask?: string; // base64 diff mask\n}\n\n/** All possible diffs for a single monitor tick. */\nexport interface MonitorDiffs {\n  markdown?: TextChange[];\n  html?: TextChange[];\n  json?: JsonChange[];\n  summary?: TextChange[];\n  links?: SetChange;\n  images?: SetChange;\n  branding?: JsonChange[];\n  screenshot?: ImageChange;\n}\n\n/**\n * Compute line-level text diffs between two strings.\n * Returns TextChange[] showing exactly which lines were added or removed.\n * No external dependencies \u2014 pure JS Myers-inspired algorithm.\n */\nexport function computeTextDiffs(previous: string, current: string): TextChange[] {\n  const prevLines = previous.split('\\n');\n  const currLines = current.split('\\n');\n  const changes: TextChange[] = [];\n\n  const prevSet = new Set(prevLines);\n  const currSet = new Set(currLines);\n\n  let lineNum = 0;\n  for (const line of prevLines) {\n    if (!currSet.has(line)) {\n      changes.push({ type: 'removed', line: lineNum, content: line });\n    }\n    lineNum++;\n  }\n  lineNum = 0;\n  for (const line of currLines) {\n    if (!prevSet.has(line)) {\n      changes.push({ type: 'added', line: lineNum, content: line });\n    }\n    lineNum++;\n  }\n\n  return changes;\n}\n\n/**\n * Compute set-based diffs for URL collections (links, images).\n */\nexport function computeSetDiffs(previous: string[], current: string[]): SetChange {\n  const prevSet = new Set(previous);\n  const currSet = new Set(current);\n  return {\n    added: current.filter(u => !prevSet.has(u)),\n    removed: previous.filter(u => !currSet.has(u)),\n  };\n}\n\n/**\n * Flatten a JSON object into dot-notation path/value pairs.\n * Used to compute JsonChange[] between two JSON extraction results.\n */\nexport function flattenJson(obj: unknown, prefix = ''): Record<string, unknown> {\n  if (typeof obj !== 'object' || obj === null) {\n    return { [prefix]: obj };\n  }\n  const result: Record<string, unknown> = {};\n  for (const [k, v] of Object.entries(obj as Record<string, unknown>)) {\n    const path = prefix ? `${prefix}.${k}` : k;\n    if (typeof v === 'object' && v !== null && !Array.isArray(v)) {\n      Object.assign(result, flattenJson(v, path));\n    } else {\n      result[path] = v;\n    }\n  }\n  return result;\n}\n\n/**\n * Compute JSON path-level diffs between two extraction results.\n */\nexport function computeJsonDiffs(previous: unknown, current: unknown): JsonChange[] {\n  const prevFlat = flattenJson(previous);\n  const currFlat = flattenJson(current);\n  const allKeys = new Set([...Object.keys(prevFlat), ...Object.keys(currFlat)]);\n  const changes: JsonChange[] = [];\n  for (const path of allKeys) {\n    const oldVal = prevFlat[path];\n    const newVal = currFlat[path];\n    if (JSON.stringify(oldVal) !== JSON.stringify(newVal)) {\n      changes.push({ path, old: oldVal, new: newVal });\n    }\n  }\n  return changes;\n}\n\n/**\n * Count total number of changes across all diff types.\n */\nexport function countMonitorDiffs(diffs: MonitorDiffs): number {\n  let count = 0;\n  if (diffs.markdown) count += diffs.markdown.length;\n  if (diffs.html) count += diffs.html.length;\n  if (diffs.json) count += diffs.json.length;\n  if (diffs.summary) count += diffs.summary.length;\n  if (diffs.links) count += diffs.links.added.length + diffs.links.removed.length;\n  if (diffs.images) count += diffs.images.added.length + diffs.images.removed.length;\n  if (diffs.branding) count += diffs.branding.length;\n  if (diffs.screenshot) count += diffs.screenshot.changed > 0 ? 1 : 0;\n  return count;\n}\n"],
+  "mappings": "AA+BO,SAAS,YAAY,MAAsB;AAChD,MAAI,KAAK;AACT,MAAI,KAAK;AACT,WAAS,IAAI,GAAG,IAAI,KAAK,QAAQ,KAAK;AACpC,UAAM,IAAI,KAAK,WAAW,CAAC;AAC3B,UAAM,KAAK,KAAK,IAAI,EAAE,IAAI,OAAO;AACjC,UAAM,KAAK,KAAK,IAAI,EAAE,IAAI,OAAO;AAAA,EACnC;AACA,UAAQ,OAAO,GAAG,SAAS,EAAE,EAAE,SAAS,GAAG,GAAG,KAAK,OAAO,GAAG,SAAS,EAAE,EAAE,SAAS,GAAG,GAAG;AAC3F;AAMO,SAAS,cAAc,MAAsB;AAClD,SAAO,KACJ,MAAM,IAAI,EACV,IAAI,OAAK,EAAE,KAAK,CAAC,EACjB,OAAO,OAAK,EAAE,SAAS,CAAC,EACxB,KAAK,IAAI;AACd;AAEO,SAAS,mBAAmB,KAAa,cAAmC;AACjF,QAAM,aAAa,cAAc,YAAY;AAC7C,SAAO;AAAA,IACL;AAAA,IACA,MAAM,YAAY,UAAU;AAAA,IAC5B,YAAY,WAAW;AAAA,IACvB,aAAY,oBAAI,KAAK,GAAE,YAAY;AAAA,EACrC;AACF;AAMO,SAAS,YACd,KACA,iBACA,aACY;AACZ,QAAM,UAAU,mBAAmB,KAAK,eAAe;AACvD,QAAM,QAAQ,YAAY,IAAI,GAAG;AAEjC,MAAI,CAAC,OAAO;AACV,WAAO,EAAE,KAAK,SAAS,MAAM,aAAa,QAAQ,MAAM,QAAQ,MAAM;AAAA,EACxE;AACA,MAAI,UAAU,QAAQ,MAAM;AAC1B,WAAO,EAAE,KAAK,SAAS,OAAO,cAAc,OAAO,aAAa,QAAQ,MAAM,QAAQ,YAAY;AAAA,EACpG;AACA,SAAO,EAAE,KAAK,SAAS,MAAM,cAAc,OAAO,aAAa,QAAQ,MAAM,QAAQ,UAAU;AACjG;AAMO,SAAS,kBAAkB,eAAyF;AACzH,QAAM,MAAM,oBAAI,IAAoB;AACpC,aAAW,KAAK,eAAe;AAC7B,QAAI,EAAE,OAAO,EAAE,YAAa,KAAI,IAAI,EAAE,KAAK,EAAE,WAAW;AAAA,EAC1D;AACA,SAAO;AACT;AASO,SAAS,kBACd,MACA,aAC2C;AAC3C,QAAM,YAAsB,CAAC;AAC7B,QAAM,SAAmB,CAAC;AAC1B,aAAW,OAAO,MAAM;AACtB,QAAI,YAAY,IAAI,GAAG,GAAG;AAGxB,gBAAU,KAAK,GAAG;AAAA,IACpB,OAAO;AACL,gBAAU,KAAK,GAAG;AAAA,IACpB;AAAA,EACF;AACA,SAAO,EAAE,WAAW,OAAO;AAC7B;AAgDO,SAAS,iBAAiB,UAAkB,SAA+B;AAChF,QAAM,YAAY,SAAS,MAAM,IAAI;AACrC,QAAM,YAAY,QAAQ,MAAM,IAAI;AACpC,QAAM,UAAwB,CAAC;AAE/B,QAAM,UAAU,IAAI,IAAI,SAAS;AACjC,QAAM,UAAU,IAAI,IAAI,SAAS;AAEjC,MAAI,UAAU;AACd,aAAW,QAAQ,WAAW;AAC5B,QAAI,CAAC,QAAQ,IAAI,IAAI,GAAG;AACtB,cAAQ,KAAK,EAAE,MAAM,WAAW,MAAM,SAAS,SAAS,KAAK,CAAC;AAAA,IAChE;AACA;AAAA,EACF;AACA,YAAU;AACV,aAAW,QAAQ,WAAW;AAC5B,QAAI,CAAC,QAAQ,IAAI,IAAI,GAAG;AACtB,cAAQ,KAAK,EAAE,MAAM,SAAS,MAAM,SAAS,SAAS,KAAK,CAAC;AAAA,IAC9D;AACA;AAAA,EACF;AAEA,SAAO;AACT;AAKO,SAAS,gBAAgB,UAAoB,SAA8B;AAChF,QAAM,UAAU,IAAI,IAAI,QAAQ;AAChC,QAAM,UAAU,IAAI,IAAI,OAAO;AAC/B,SAAO;AAAA,IACL,OAAO,QAAQ,OAAO,OAAK,CAAC,QAAQ,IAAI,CAAC,CAAC;AAAA,IAC1C,SAAS,SAAS,OAAO,OAAK,CAAC,QAAQ,IAAI,CAAC,CAAC;AAAA,EAC/C;AACF;AAMO,SAAS,YAAY,KAAc,SAAS,IAA6B;AAC9E,MAAI,OAAO,QAAQ,YAAY,QAAQ,MAAM;AAC3C,WAAO,EAAE,CAAC,MAAM,GAAG,IAAI;AAAA,EACzB;AACA,QAAM,SAAkC,CAAC;AACzC,aAAW,CAAC,GAAG,CAAC,KAAK,OAAO,QAAQ,GAA8B,GAAG;AACnE,UAAM,OAAO,SAAS,GAAG,MAAM,IAAI,CAAC,KAAK;AACzC,QAAI,OAAO,MAAM,YAAY,MAAM,QAAQ,CAAC,MAAM,QAAQ,CAAC,GAAG;AAC5D,aAAO,OAAO,QAAQ,YAAY,GAAG,IAAI,CAAC;AAAA,IAC5C,OAAO;AACL,aAAO,IAAI,IAAI;AAAA,IACjB;AAAA,EACF;AACA,SAAO;AACT;AAKO,SAAS,iBAAiB,UAAmB,SAAgC;AAClF,QAAM,WAAW,YAAY,QAAQ;AACrC,QAAM,WAAW,YAAY,OAAO;AACpC,QAAM,UAAU,oBAAI,IAAI,CAAC,GAAG,OAAO,KAAK,QAAQ,GAAG,GAAG,OAAO,KAAK,QAAQ,CAAC,CAAC;AAC5E,QAAM,UAAwB,CAAC;AAC/B,aAAW,QAAQ,SAAS;AAC1B,UAAM,SAAS,SAAS,IAAI;AAC5B,UAAM,SAAS,SAAS,IAAI;AAC5B,QAAI,KAAK,UAAU,MAAM,MAAM,KAAK,UAAU,MAAM,GAAG;AACrD,cAAQ,KAAK,EAAE,MAAM,KAAK,QAAQ,KAAK,OAAO,CAAC;AAAA,IACjD;AAAA,EACF;AACA,SAAO;AACT;AAKO,SAAS,kBAAkB,OAA6B;AAC7D,MAAI,QAAQ;AACZ,MAAI,MAAM,SAAU,UAAS,MAAM,SAAS;AAC5C,MAAI,MAAM,KAAM,UAAS,MAAM,KAAK;AACpC,MAAI,MAAM,KAAM,UAAS,MAAM,KAAK;AACpC,MAAI,MAAM,QAAS,UAAS,MAAM,QAAQ;AAC1C,MAAI,MAAM,MAAO,UAAS,MAAM,MAAM,MAAM,SAAS,MAAM,MAAM,QAAQ;AACzE,MAAI,MAAM,OAAQ,UAAS,MAAM,OAAO,MAAM,SAAS,MAAM,OAAO,QAAQ;AAC5E,MAAI,MAAM,SAAU,UAAS,MAAM,SAAS;AAC5C,MAAI,MAAM,WAAY,UAAS,MAAM,WAAW,UAAU,IAAI,IAAI;AAClE,SAAO;AACT;",
+  "names": []
+}

package/dist/index.d.ts

@@ -0,0 +1,52 @@
+export * from './crawler.js';
+export * from './visual-diff-verdict.js';
+export * from './auth.js';
+export * from './screen-reader.js';
+export * from './job-queue.js';
+export { startSqsWorker, getDlqStats, redriveFromDlq } from './sqs-queue.js';
+export * from './session-capture.js';
+export * from './session-export.js';
+export * from './test-executor.js';
+export { startLiveView, getFrame, liveViewNavigate, completeLiveView, closeLiveView, sendInput } from './live-view.js';
+export { waitForEmailOTP, getInboxEmail } from './email-otp.js';
+export { getSsoToken, getSsoClientCredentialsToken } from './sso-token.js';
+export type { SsoConfig, SsoProvider } from './sso-token.js';
+export * from './engine-router.js';
+export * from './profiles/domain-profile.js';
+export * from './policies/crawl-safety.js';
+export * from './extractors/artifact-envelope.js';
+export * from './network-observer.js';
+export { createChromeTlsAgent, probeTlsFingerprint, chromeTlsAgent, CHROME_TLS_HEADERS } from './tls-fingerprint.js';
+export type { TlsFingerprintOptions } from './tls-fingerprint.js';
+export { prepareExtensionProfile, adBlockerProfile } from './browser-extension.js';
+export type { ExtensionSource, ExtensionProfile } from './browser-extension.js';
+export { createRedisQueue, maybeCreateRedisQueue, isRedisQueueEnabled } from './redis-queue.js';
+export type { RedisQueueJob, RedisQueueClient } from './redis-queue.js';
+export { discoverProjectUrls, discoverSitemapUrls } from './sitemap.js';
+export type { UrlCandidate } from './sitemap.js';
+export { httpCrawl } from './http-engine.js';
+export { extractMarkdown } from './markdown-extract.js';
+export type { ExtractedMarkdown } from './markdown-extract.js';
+export type { HttpScreen, HttpCrawlOpts, HttpCrawlResult, HttpOutputFormat } from './http-engine.js';
+export { TieredProxyManager } from './proxy.js';
+export type { ProxyTier, ProxyTierConfig } from './proxy.js';
+export { RedisCrawlFrontier, isRedisFrontierEnabled } from './redis-frontier.js';
+export { startSpan, endSpan, withSpan, increment, flushMetrics } from './telemetry.js';
+export type { Span, FinishedSpan } from './telemetry.js';
+export { analyzeFailure } from './failure-analysis-agent.js';
+export type { FailureContext } from './failure-analysis-agent.js';
+export { elementToPlaywrightLocator, elementsToPlaywrightLocators, formatLocatorsForPrompt, detectAuthWallLocators } from './locator-extractor.js';
+export type { PlaywrightLocator } from './locator-extractor.js';
+export * from './dropdown-enumerator.js';
+export * from './crud-interaction.js';
+export * from './filter-pagination-crawler.js';
+export * from './multi-context-handler.js';
+export * from './advanced-ui-probes.js';
+export * from './state-graph.js';
+export * from './nested-window-crawler.js';
+export * from './shadow-dom-crawler.js';
+export * from './network-action-correlator.js';
+export * from './crawl-session-manager.js';
+export * from './accessibility-tree-crawler.js';
+export * from './performance-metrics.js';
+export * from './service-worker-interceptor.js';

package/dist/index.js

@@ -0,0 +1,83 @@
+export * from "./crawler.js";
+export * from "./visual-diff-verdict.js";
+export * from "./auth.js";
+export * from "./screen-reader.js";
+export * from "./job-queue.js";
+import { startSqsWorker, getDlqStats, redriveFromDlq } from "./sqs-queue.js";
+export * from "./session-capture.js";
+export * from "./session-export.js";
+export * from "./test-executor.js";
+import { startLiveView, getFrame, liveViewNavigate, completeLiveView, closeLiveView, sendInput } from "./live-view.js";
+import { waitForEmailOTP, getInboxEmail } from "./email-otp.js";
+import { getSsoToken, getSsoClientCredentialsToken } from "./sso-token.js";
+export * from "./engine-router.js";
+export * from "./profiles/domain-profile.js";
+export * from "./policies/crawl-safety.js";
+export * from "./extractors/artifact-envelope.js";
+export * from "./network-observer.js";
+import { createChromeTlsAgent, probeTlsFingerprint, chromeTlsAgent, CHROME_TLS_HEADERS } from "./tls-fingerprint.js";
+import { prepareExtensionProfile, adBlockerProfile } from "./browser-extension.js";
+import { createRedisQueue, maybeCreateRedisQueue, isRedisQueueEnabled } from "./redis-queue.js";
+import { discoverProjectUrls, discoverSitemapUrls } from "./sitemap.js";
+import { httpCrawl } from "./http-engine.js";
+import { extractMarkdown } from "./markdown-extract.js";
+import { TieredProxyManager } from "./proxy.js";
+import { RedisCrawlFrontier, isRedisFrontierEnabled } from "./redis-frontier.js";
+import { startSpan, endSpan, withSpan, increment, flushMetrics } from "./telemetry.js";
+import { analyzeFailure } from "./failure-analysis-agent.js";
+import { elementToPlaywrightLocator, elementsToPlaywrightLocators, formatLocatorsForPrompt, detectAuthWallLocators } from "./locator-extractor.js";
+export * from "./dropdown-enumerator.js";
+export * from "./crud-interaction.js";
+export * from "./filter-pagination-crawler.js";
+export * from "./multi-context-handler.js";
+export * from "./advanced-ui-probes.js";
+export * from "./state-graph.js";
+export * from "./nested-window-crawler.js";
+export * from "./shadow-dom-crawler.js";
+export * from "./network-action-correlator.js";
+export * from "./crawl-session-manager.js";
+export * from "./accessibility-tree-crawler.js";
+export * from "./performance-metrics.js";
+export * from "./service-worker-interceptor.js";
+export {
+  CHROME_TLS_HEADERS,
+  RedisCrawlFrontier,
+  TieredProxyManager,
+  adBlockerProfile,
+  analyzeFailure,
+  chromeTlsAgent,
+  closeLiveView,
+  completeLiveView,
+  createChromeTlsAgent,
+  createRedisQueue,
+  detectAuthWallLocators,
+  discoverProjectUrls,
+  discoverSitemapUrls,
+  elementToPlaywrightLocator,
+  elementsToPlaywrightLocators,
+  endSpan,
+  extractMarkdown,
+  flushMetrics,
+  formatLocatorsForPrompt,
+  getDlqStats,
+  getFrame,
+  getInboxEmail,
+  getSsoClientCredentialsToken,
+  getSsoToken,
+  httpCrawl,
+  increment,
+  isRedisFrontierEnabled,
+  isRedisQueueEnabled,
+  liveViewNavigate,
+  maybeCreateRedisQueue,
+  prepareExtensionProfile,
+  probeTlsFingerprint,
+  redriveFromDlq,
+  sendInput,
+  startLiveView,
+  startSpan,
+  startSqsWorker,
+  waitForEmailOTP,
+  withSpan
+};
+//# sourceMappingURL=index.js.map

package/dist/index.js.map

@@ -0,0 +1,7 @@
+{
+  "version": 3,
+  "sources": ["../src/index.ts"],
+  "sourcesContent": ["export * from './crawler.js';\nexport * from './visual-diff-verdict.js';\nexport * from './auth.js';\nexport * from './screen-reader.js';\nexport * from './job-queue.js';\nexport { startSqsWorker, getDlqStats, redriveFromDlq } from './sqs-queue.js';\nexport * from './session-capture.js';\nexport * from './session-export.js';\nexport * from './test-executor.js';\nexport { startLiveView, getFrame, liveViewNavigate, completeLiveView, closeLiveView, sendInput } from './live-view.js';\nexport { waitForEmailOTP, getInboxEmail } from './email-otp.js';\nexport { getSsoToken, getSsoClientCredentialsToken } from './sso-token.js';\nexport type { SsoConfig, SsoProvider } from './sso-token.js';\nexport * from './engine-router.js';\nexport * from './profiles/domain-profile.js';\nexport * from './policies/crawl-safety.js';\nexport * from './extractors/artifact-envelope.js';\nexport * from './network-observer.js';\nexport { createChromeTlsAgent, probeTlsFingerprint, chromeTlsAgent, CHROME_TLS_HEADERS } from './tls-fingerprint.js';\nexport type { TlsFingerprintOptions } from './tls-fingerprint.js';\nexport { prepareExtensionProfile, adBlockerProfile } from './browser-extension.js';\nexport type { ExtensionSource, ExtensionProfile } from './browser-extension.js';\nexport { createRedisQueue, maybeCreateRedisQueue, isRedisQueueEnabled } from './redis-queue.js';\nexport type { RedisQueueJob, RedisQueueClient } from './redis-queue.js';\nexport { discoverProjectUrls, discoverSitemapUrls } from './sitemap.js';\nexport type { UrlCandidate } from './sitemap.js';\nexport { httpCrawl } from './http-engine.js';\nexport { extractMarkdown } from './markdown-extract.js';\nexport type { ExtractedMarkdown } from './markdown-extract.js';\nexport type { HttpScreen, HttpCrawlOpts, HttpCrawlResult, HttpOutputFormat } from './http-engine.js';\nexport { TieredProxyManager } from './proxy.js';\nexport type { ProxyTier, ProxyTierConfig } from './proxy.js';\nexport { RedisCrawlFrontier, isRedisFrontierEnabled } from './redis-frontier.js';\nexport { startSpan, endSpan, withSpan, increment, flushMetrics } from './telemetry.js';\nexport type { Span, FinishedSpan } from './telemetry.js';\nexport { analyzeFailure } from './failure-analysis-agent.js';\nexport type { FailureContext } from './failure-analysis-agent.js';\nexport { elementToPlaywrightLocator, elementsToPlaywrightLocators, formatLocatorsForPrompt, detectAuthWallLocators } from './locator-extractor.js';\nexport type { PlaywrightLocator } from './locator-extractor.js';\nexport * from './dropdown-enumerator.js';\nexport * from './crud-interaction.js';\nexport * from './filter-pagination-crawler.js';\nexport * from './multi-context-handler.js';\nexport * from './advanced-ui-probes.js';\nexport * from './state-graph.js';\nexport * from './nested-window-crawler.js';\nexport * from './shadow-dom-crawler.js';\nexport * from './network-action-correlator.js';\nexport * from './crawl-session-manager.js';\nexport * from './accessibility-tree-crawler.js';\nexport * from './performance-metrics.js';\nexport * from './service-worker-interceptor.js';\n"],
+  "mappings": "AAAA,cAAc;AACd,cAAc;AACd,cAAc;AACd,cAAc;AACd,cAAc;AACd,SAAS,gBAAgB,aAAa,sBAAsB;AAC5D,cAAc;AACd,cAAc;AACd,cAAc;AACd,SAAS,eAAe,UAAU,kBAAkB,kBAAkB,eAAe,iBAAiB;AACtG,SAAS,iBAAiB,qBAAqB;AAC/C,SAAS,aAAa,oCAAoC;AAE1D,cAAc;AACd,cAAc;AACd,cAAc;AACd,cAAc;AACd,cAAc;AACd,SAAS,sBAAsB,qBAAqB,gBAAgB,0BAA0B;AAE9F,SAAS,yBAAyB,wBAAwB;AAE1D,SAAS,kBAAkB,uBAAuB,2BAA2B;AAE7E,SAAS,qBAAqB,2BAA2B;AAEzD,SAAS,iBAAiB;AAC1B,SAAS,uBAAuB;AAGhC,SAAS,0BAA0B;AAEnC,SAAS,oBAAoB,8BAA8B;AAC3D,SAAS,WAAW,SAAS,UAAU,WAAW,oBAAoB;AAEtE,SAAS,sBAAsB;AAE/B,SAAS,4BAA4B,8BAA8B,yBAAyB,8BAA8B;AAE1H,cAAc;AACd,cAAc;AACd,cAAc;AACd,cAAc;AACd,cAAc;AACd,cAAc;AACd,cAAc;AACd,cAAc;AACd,cAAc;AACd,cAAc;AACd,cAAc;AACd,cAAc;AACd,cAAc;",
+  "names": []
+}

package/dist/input-fuzzer.d.ts

@@ -0,0 +1,30 @@
+/**
+ * Input fuzzer — generate test data for form fields.
+ *
+ * For each field type, produces a set of test values covering:
+ * - Happy path (valid, normal input)
+ * - Boundary (min/max length, min/max value)
+ * - Edge cases (empty, very long, special chars, SQL injection probe, XSS probe)
+ * - Type-invalid (wrong type for the field)
+ *
+ * This feeds ZeTa's test generator to automatically create form validation tests.
+ * Values are safe probes (NOT actual exploits) — they test whether the app
+ * properly rejects/accepts boundary inputs.
+ */
+import type { FormField } from './form-detector.js';
+export type InputCase = {
+    label: string;
+    value: string;
+    expectedValid: boolean;
+    category: 'happy' | 'boundary' | 'edge' | 'invalid';
+};
+export declare function generateFieldCases(field: FormField): InputCase[];
+export interface FuzzedForm {
+    formIndex: number;
+    purpose: string;
+    fieldCases: Array<{
+        field: FormField;
+        cases: InputCase[];
+    }>;
+}
+export declare function fuzzForm(form: import('./form-detector.js').DetectedForm): FuzzedForm;

package/dist/input-fuzzer.js

@@ -0,0 +1,105 @@
+const LOREM = "Lorem ipsum dolor sit amet consectetur adipiscing elit";
+function generateEmailCases() {
+  return [
+    { label: "valid email", value: "test@example.com", expectedValid: true, category: "happy" },
+    { label: "email with plus", value: "test+tag@example.com", expectedValid: true, category: "happy" },
+    { label: "missing @", value: "notanemail.com", expectedValid: false, category: "invalid" },
+    { label: "missing domain", value: "test@", expectedValid: false, category: "invalid" },
+    { label: "empty", value: "", expectedValid: false, category: "edge" },
+    { label: "very long email", value: "a".repeat(100) + "@example.com", expectedValid: false, category: "boundary" },
+    { label: "special chars", value: "test<script>@example.com", expectedValid: false, category: "edge" }
+  ];
+}
+function generatePasswordCases(minLength, maxLength) {
+  const min = minLength ?? 8;
+  const max = maxLength ?? 128;
+  return [
+    { label: "valid password", value: "ValidP@ss1", expectedValid: true, category: "happy" },
+    { label: "minimum length", value: "A".repeat(min), expectedValid: true, category: "boundary" },
+    { label: "below minimum", value: "A".repeat(Math.max(1, min - 1)), expectedValid: false, category: "boundary" },
+    { label: "max length", value: "Aa1!".repeat(Math.ceil(max / 4)).slice(0, max), expectedValid: true, category: "boundary" },
+    { label: "empty", value: "", expectedValid: false, category: "edge" },
+    { label: "spaces only", value: "        ", expectedValid: false, category: "edge" }
+  ];
+}
+function generateTextCases(minLength, maxLength, required) {
+  const cases = [
+    { label: "typical text", value: "Test Input Value", expectedValid: true, category: "happy" },
+    { label: "single char", value: "A", expectedValid: !minLength || minLength <= 1, category: "boundary" },
+    { label: "empty", value: "", expectedValid: !required, category: "edge" },
+    { label: "spaces only", value: "   ", expectedValid: !required, category: "edge" },
+    { label: "very long text", value: LOREM.repeat(20), expectedValid: !maxLength || maxLength > LOREM.length * 20, category: "boundary" },
+    { label: "unicode", value: "\u65E5\u672C\u8A9E\u30C6\u30B9\u30C8 \xDCn\xEFc\xF6d\xE9 Emojitest \u{1F389}", expectedValid: true, category: "edge" },
+    { label: "HTML special chars", value: "<script>alert(1)</script>", expectedValid: true, category: "edge" },
+    { label: "SQL probe", value: "' OR '1'='1", expectedValid: true, category: "edge" }
+  ];
+  if (minLength) cases.push({ label: `below min (${minLength})`, value: "x".repeat(Math.max(0, minLength - 1)), expectedValid: false, category: "boundary" });
+  if (maxLength) cases.push({ label: `above max (${maxLength})`, value: "x".repeat(maxLength + 1), expectedValid: false, category: "boundary" });
+  return cases;
+}
+function generateNumberCases(min, max) {
+  const minN = min !== void 0 ? parseFloat(min) : void 0;
+  const maxN = max !== void 0 ? parseFloat(max) : void 0;
+  return [
+    { label: "valid number", value: "42", expectedValid: true, category: "happy" },
+    { label: "zero", value: "0", expectedValid: minN === void 0 || minN <= 0, category: "boundary" },
+    { label: "negative", value: "-1", expectedValid: minN === void 0 || minN < 0, category: "boundary" },
+    { label: "decimal", value: "3.14", expectedValid: true, category: "happy" },
+    { label: "text as number", value: "notanumber", expectedValid: false, category: "invalid" },
+    { label: "empty", value: "", expectedValid: false, category: "edge" },
+    ...minN !== void 0 ? [{ label: `below min (${min})`, value: String(minN - 1), expectedValid: false, category: "boundary" }] : [],
+    ...maxN !== void 0 ? [{ label: `above max (${max})`, value: String(maxN + 1), expectedValid: false, category: "boundary" }] : []
+  ];
+}
+function generateSelectCases(options) {
+  if (!options?.length) return [{ label: "first option", value: "", expectedValid: true, category: "happy" }];
+  return [
+    { label: `first option (${options[0]})`, value: options[0], expectedValid: true, category: "happy" },
+    { label: "invalid option", value: "__INVALID_OPTION__", expectedValid: false, category: "invalid" },
+    { label: "empty selection", value: "", expectedValid: false, category: "edge" },
+    ...options.length > 1 ? [{ label: `last option (${options[options.length - 1]})`, value: options[options.length - 1], expectedValid: true, category: "happy" }] : []
+  ];
+}
+function generateFieldCases(field) {
+  switch (field.type) {
+    case "email":
+      return generateEmailCases();
+    case "password":
+      return generatePasswordCases(field.minLength, field.maxLength);
+    case "number":
+      return generateNumberCases(field.min, field.max);
+    case "select":
+      return generateSelectCases(field.options);
+    case "checkbox":
+      return [
+        { label: "checked", value: "true", expectedValid: true, category: "happy" },
+        { label: "unchecked", value: "false", expectedValid: !field.required, category: field.required ? "invalid" : "happy" }
+      ];
+    case "date":
+      return [
+        { label: "valid date", value: "2026-01-15", expectedValid: true, category: "happy" },
+        { label: "invalid date", value: "9999-99-99", expectedValid: false, category: "invalid" },
+        { label: "empty", value: "", expectedValid: !field.required, category: "edge" }
+      ];
+    case "url":
+      return [
+        { label: "valid URL", value: "https://example.com", expectedValid: true, category: "happy" },
+        { label: "missing protocol", value: "example.com", expectedValid: false, category: "invalid" },
+        { label: "empty", value: "", expectedValid: !field.required, category: "edge" }
+      ];
+    default:
+      return generateTextCases(field.minLength, field.maxLength, field.required);
+  }
+}
+function fuzzForm(form) {
+  return {
+    formIndex: form.index,
+    purpose: form.purpose,
+    fieldCases: form.fields.map((field) => ({ field, cases: generateFieldCases(field) }))
+  };
+}
+export {
+  fuzzForm,
+  generateFieldCases
+};
+//# sourceMappingURL=input-fuzzer.js.map

package/dist/input-fuzzer.js.map

@@ -0,0 +1,7 @@
+{
+  "version": 3,
+  "sources": ["../src/input-fuzzer.ts"],
+  "sourcesContent": ["/**\n * Input fuzzer \u2014 generate test data for form fields.\n *\n * For each field type, produces a set of test values covering:\n * - Happy path (valid, normal input)\n * - Boundary (min/max length, min/max value)\n * - Edge cases (empty, very long, special chars, SQL injection probe, XSS probe)\n * - Type-invalid (wrong type for the field)\n *\n * This feeds ZeTa's test generator to automatically create form validation tests.\n * Values are safe probes (NOT actual exploits) \u2014 they test whether the app\n * properly rejects/accepts boundary inputs.\n */\n\nimport type { FormField } from './form-detector.js';\n\nexport type InputCase = {\n  label: string;          // human-readable test case name\n  value: string;\n  expectedValid: boolean; // true if input should be accepted\n  category: 'happy' | 'boundary' | 'edge' | 'invalid';\n};\n\nconst LOREM = 'Lorem ipsum dolor sit amet consectetur adipiscing elit';\n\nfunction generateEmailCases(): InputCase[] {\n  return [\n    { label: 'valid email', value: 'test@example.com', expectedValid: true, category: 'happy' },\n    { label: 'email with plus', value: 'test+tag@example.com', expectedValid: true, category: 'happy' },\n    { label: 'missing @', value: 'notanemail.com', expectedValid: false, category: 'invalid' },\n    { label: 'missing domain', value: 'test@', expectedValid: false, category: 'invalid' },\n    { label: 'empty', value: '', expectedValid: false, category: 'edge' },\n    { label: 'very long email', value: 'a'.repeat(100) + '@example.com', expectedValid: false, category: 'boundary' },\n    { label: 'special chars', value: 'test<script>@example.com', expectedValid: false, category: 'edge' },\n  ];\n}\n\nfunction generatePasswordCases(minLength?: number, maxLength?: number): InputCase[] {\n  const min = minLength ?? 8;\n  const max = maxLength ?? 128;\n  return [\n    { label: 'valid password', value: 'ValidP@ss1', expectedValid: true, category: 'happy' },\n    { label: 'minimum length', value: 'A'.repeat(min), expectedValid: true, category: 'boundary' },\n    { label: 'below minimum', value: 'A'.repeat(Math.max(1, min - 1)), expectedValid: false, category: 'boundary' },\n    { label: 'max length', value: 'Aa1!'.repeat(Math.ceil(max / 4)).slice(0, max), expectedValid: true, category: 'boundary' },\n    { label: 'empty', value: '', expectedValid: false, category: 'edge' },\n    { label: 'spaces only', value: '        ', expectedValid: false, category: 'edge' },\n  ];\n}\n\nfunction generateTextCases(minLength?: number, maxLength?: number, required?: boolean): InputCase[] {\n  const cases: InputCase[] = [\n    { label: 'typical text', value: 'Test Input Value', expectedValid: true, category: 'happy' },\n    { label: 'single char', value: 'A', expectedValid: !minLength || minLength <= 1, category: 'boundary' },\n    { label: 'empty', value: '', expectedValid: !required, category: 'edge' },\n    { label: 'spaces only', value: '   ', expectedValid: !required, category: 'edge' },\n    { label: 'very long text', value: LOREM.repeat(20), expectedValid: !maxLength || maxLength > LOREM.length * 20, category: 'boundary' },\n    { label: 'unicode', value: '\u65E5\u672C\u8A9E\u30C6\u30B9\u30C8 \u00DCn\u00EFc\u00F6d\u00E9 Emojitest \uD83C\uDF89', expectedValid: true, category: 'edge' },\n    { label: 'HTML special chars', value: '<script>alert(1)</script>', expectedValid: true, category: 'edge' },\n    { label: 'SQL probe', value: \"' OR '1'='1\", expectedValid: true, category: 'edge' },\n  ];\n  if (minLength) cases.push({ label: `below min (${minLength})`, value: 'x'.repeat(Math.max(0, minLength - 1)), expectedValid: false, category: 'boundary' });\n  if (maxLength) cases.push({ label: `above max (${maxLength})`, value: 'x'.repeat(maxLength + 1), expectedValid: false, category: 'boundary' });\n  return cases;\n}\n\nfunction generateNumberCases(min?: string, max?: string): InputCase[] {\n  const minN = min !== undefined ? parseFloat(min) : undefined;\n  const maxN = max !== undefined ? parseFloat(max) : undefined;\n  return [\n    { label: 'valid number', value: '42', expectedValid: true, category: 'happy' },\n    { label: 'zero', value: '0', expectedValid: minN === undefined || minN <= 0, category: 'boundary' },\n    { label: 'negative', value: '-1', expectedValid: minN === undefined || minN < 0, category: 'boundary' },\n    { label: 'decimal', value: '3.14', expectedValid: true, category: 'happy' },\n    { label: 'text as number', value: 'notanumber', expectedValid: false, category: 'invalid' },\n    { label: 'empty', value: '', expectedValid: false, category: 'edge' },\n    ...(minN !== undefined ? [{ label: `below min (${min})`, value: String(minN - 1), expectedValid: false, category: 'boundary' as const }] : []),\n    ...(maxN !== undefined ? [{ label: `above max (${max})`, value: String(maxN + 1), expectedValid: false, category: 'boundary' as const }] : []),\n  ];\n}\n\nfunction generateSelectCases(options?: string[]): InputCase[] {\n  if (!options?.length) return [{ label: 'first option', value: '', expectedValid: true, category: 'happy' }];\n  return [\n    { label: `first option (${options[0]})`, value: options[0], expectedValid: true, category: 'happy' },\n    { label: 'invalid option', value: '__INVALID_OPTION__', expectedValid: false, category: 'invalid' },\n    { label: 'empty selection', value: '', expectedValid: false, category: 'edge' },\n    ...(options.length > 1 ? [{ label: `last option (${options[options.length - 1]})`, value: options[options.length - 1], expectedValid: true, category: 'happy' as const }] : []),\n  ];\n}\n\nexport function generateFieldCases(field: FormField): InputCase[] {\n  switch (field.type) {\n    case 'email':    return generateEmailCases();\n    case 'password': return generatePasswordCases(field.minLength, field.maxLength);\n    case 'number':   return generateNumberCases(field.min, field.max);\n    case 'select':   return generateSelectCases(field.options);\n    case 'checkbox': return [\n      { label: 'checked', value: 'true', expectedValid: true, category: 'happy' },\n      { label: 'unchecked', value: 'false', expectedValid: !field.required, category: field.required ? 'invalid' : 'happy' },\n    ];\n    case 'date':     return [\n      { label: 'valid date', value: '2026-01-15', expectedValid: true, category: 'happy' },\n      { label: 'invalid date', value: '9999-99-99', expectedValid: false, category: 'invalid' },\n      { label: 'empty', value: '', expectedValid: !field.required, category: 'edge' },\n    ];\n    case 'url':      return [\n      { label: 'valid URL', value: 'https://example.com', expectedValid: true, category: 'happy' },\n      { label: 'missing protocol', value: 'example.com', expectedValid: false, category: 'invalid' },\n      { label: 'empty', value: '', expectedValid: !field.required, category: 'edge' },\n    ];\n    default:\n      return generateTextCases(field.minLength, field.maxLength, field.required);\n  }\n}\n\nexport interface FuzzedForm {\n  formIndex: number;\n  purpose: string;\n  fieldCases: Array<{ field: FormField; cases: InputCase[] }>;\n}\n\nexport function fuzzForm(form: import('./form-detector.js').DetectedForm): FuzzedForm {\n  return {\n    formIndex: form.index,\n    purpose: form.purpose,\n    fieldCases: form.fields.map(field => ({ field, cases: generateFieldCases(field) })),\n  };\n}\n"],
+  "mappings": "AAuBA,MAAM,QAAQ;AAEd,SAAS,qBAAkC;AACzC,SAAO;AAAA,IACL,EAAE,OAAO,eAAe,OAAO,oBAAoB,eAAe,MAAM,UAAU,QAAQ;AAAA,IAC1F,EAAE,OAAO,mBAAmB,OAAO,wBAAwB,eAAe,MAAM,UAAU,QAAQ;AAAA,IAClG,EAAE,OAAO,aAAa,OAAO,kBAAkB,eAAe,OAAO,UAAU,UAAU;AAAA,IACzF,EAAE,OAAO,kBAAkB,OAAO,SAAS,eAAe,OAAO,UAAU,UAAU;AAAA,IACrF,EAAE,OAAO,SAAS,OAAO,IAAI,eAAe,OAAO,UAAU,OAAO;AAAA,IACpE,EAAE,OAAO,mBAAmB,OAAO,IAAI,OAAO,GAAG,IAAI,gBAAgB,eAAe,OAAO,UAAU,WAAW;AAAA,IAChH,EAAE,OAAO,iBAAiB,OAAO,4BAA4B,eAAe,OAAO,UAAU,OAAO;AAAA,EACtG;AACF;AAEA,SAAS,sBAAsB,WAAoB,WAAiC;AAClF,QAAM,MAAM,aAAa;AACzB,QAAM,MAAM,aAAa;AACzB,SAAO;AAAA,IACL,EAAE,OAAO,kBAAkB,OAAO,cAAc,eAAe,MAAM,UAAU,QAAQ;AAAA,IACvF,EAAE,OAAO,kBAAkB,OAAO,IAAI,OAAO,GAAG,GAAG,eAAe,MAAM,UAAU,WAAW;AAAA,IAC7F,EAAE,OAAO,iBAAiB,OAAO,IAAI,OAAO,KAAK,IAAI,GAAG,MAAM,CAAC,CAAC,GAAG,eAAe,OAAO,UAAU,WAAW;AAAA,IAC9G,EAAE,OAAO,cAAc,OAAO,OAAO,OAAO,KAAK,KAAK,MAAM,CAAC,CAAC,EAAE,MAAM,GAAG,GAAG,GAAG,eAAe,MAAM,UAAU,WAAW;AAAA,IACzH,EAAE,OAAO,SAAS,OAAO,IAAI,eAAe,OAAO,UAAU,OAAO;AAAA,IACpE,EAAE,OAAO,eAAe,OAAO,YAAY,eAAe,OAAO,UAAU,OAAO;AAAA,EACpF;AACF;AAEA,SAAS,kBAAkB,WAAoB,WAAoB,UAAiC;AAClG,QAAM,QAAqB;AAAA,IACzB,EAAE,OAAO,gBAAgB,OAAO,oBAAoB,eAAe,MAAM,UAAU,QAAQ;AAAA,IAC3F,EAAE,OAAO,eAAe,OAAO,KAAK,eAAe,CAAC,aAAa,aAAa,GAAG,UAAU,WAAW;AAAA,IACtG,EAAE,OAAO,SAAS,OAAO,IAAI,eAAe,CAAC,UAAU,UAAU,OAAO;AAAA,IACxE,EAAE,OAAO,eAAe,OAAO,OAAO,eAAe,CAAC,UAAU,UAAU,OAAO;AAAA,IACjF,EAAE,OAAO,kBAAkB,OAAO,MAAM,OAAO,EAAE,GAAG,eAAe,CAAC,aAAa,YAAY,MAAM,SAAS,IAAI,UAAU,WAAW;AAAA,IACrI,EAAE,OAAO,WAAW,OAAO,gFAA+B,eAAe,MAAM,UAAU,OAAO;AAAA,IAChG,EAAE,OAAO,sBAAsB,OAAO,6BAA6B,eAAe,MAAM,UAAU,OAAO;AAAA,IACzG,EAAE,OAAO,aAAa,OAAO,eAAe,eAAe,MAAM,UAAU,OAAO;AAAA,EACpF;AACA,MAAI,UAAW,OAAM,KAAK,EAAE,OAAO,cAAc,SAAS,KAAK,OAAO,IAAI,OAAO,KAAK,IAAI,GAAG,YAAY,CAAC,CAAC,GAAG,eAAe,OAAO,UAAU,WAAW,CAAC;AAC1J,MAAI,UAAW,OAAM,KAAK,EAAE,OAAO,cAAc,SAAS,KAAK,OAAO,IAAI,OAAO,YAAY,CAAC,GAAG,eAAe,OAAO,UAAU,WAAW,CAAC;AAC7I,SAAO;AACT;AAEA,SAAS,oBAAoB,KAAc,KAA2B;AACpE,QAAM,OAAO,QAAQ,SAAY,WAAW,GAAG,IAAI;AACnD,QAAM,OAAO,QAAQ,SAAY,WAAW,GAAG,IAAI;AACnD,SAAO;AAAA,IACL,EAAE,OAAO,gBAAgB,OAAO,MAAM,eAAe,MAAM,UAAU,QAAQ;AAAA,IAC7E,EAAE,OAAO,QAAQ,OAAO,KAAK,eAAe,SAAS,UAAa,QAAQ,GAAG,UAAU,WAAW;AAAA,IAClG,EAAE,OAAO,YAAY,OAAO,MAAM,eAAe,SAAS,UAAa,OAAO,GAAG,UAAU,WAAW;AAAA,IACtG,EAAE,OAAO,WAAW,OAAO,QAAQ,eAAe,MAAM,UAAU,QAAQ;AAAA,IAC1E,EAAE,OAAO,kBAAkB,OAAO,cAAc,eAAe,OAAO,UAAU,UAAU;AAAA,IAC1F,EAAE,OAAO,SAAS,OAAO,IAAI,eAAe,OAAO,UAAU,OAAO;AAAA,IACpE,GAAI,SAAS,SAAY,CAAC,EAAE,OAAO,cAAc,GAAG,KAAK,OAAO,OAAO,OAAO,CAAC,GAAG,eAAe,OAAO,UAAU,WAAoB,CAAC,IAAI,CAAC;AAAA,IAC5I,GAAI,SAAS,SAAY,CAAC,EAAE,OAAO,cAAc,GAAG,KAAK,OAAO,OAAO,OAAO,CAAC,GAAG,eAAe,OAAO,UAAU,WAAoB,CAAC,IAAI,CAAC;AAAA,EAC9I;AACF;AAEA,SAAS,oBAAoB,SAAiC;AAC5D,MAAI,CAAC,SAAS,OAAQ,QAAO,CAAC,EAAE,OAAO,gBAAgB,OAAO,IAAI,eAAe,MAAM,UAAU,QAAQ,CAAC;AAC1G,SAAO;AAAA,IACL,EAAE,OAAO,iBAAiB,QAAQ,CAAC,CAAC,KAAK,OAAO,QAAQ,CAAC,GAAG,eAAe,MAAM,UAAU,QAAQ;AAAA,IACnG,EAAE,OAAO,kBAAkB,OAAO,sBAAsB,eAAe,OAAO,UAAU,UAAU;AAAA,IAClG,EAAE,OAAO,mBAAmB,OAAO,IAAI,eAAe,OAAO,UAAU,OAAO;AAAA,IAC9E,GAAI,QAAQ,SAAS,IAAI,CAAC,EAAE,OAAO,gBAAgB,QAAQ,QAAQ,SAAS,CAAC,CAAC,KAAK,OAAO,QAAQ,QAAQ,SAAS,CAAC,GAAG,eAAe,MAAM,UAAU,QAAiB,CAAC,IAAI,CAAC;AAAA,EAC/K;AACF;AAEO,SAAS,mBAAmB,OAA+B;AAChE,UAAQ,MAAM,MAAM;AAAA,IAClB,KAAK;AAAY,aAAO,mBAAmB;AAAA,IAC3C,KAAK;AAAY,aAAO,sBAAsB,MAAM,WAAW,MAAM,SAAS;AAAA,IAC9E,KAAK;AAAY,aAAO,oBAAoB,MAAM,KAAK,MAAM,GAAG;AAAA,IAChE,KAAK;AAAY,aAAO,oBAAoB,MAAM,OAAO;AAAA,IACzD,KAAK;AAAY,aAAO;AAAA,QACtB,EAAE,OAAO,WAAW,OAAO,QAAQ,eAAe,MAAM,UAAU,QAAQ;AAAA,QAC1E,EAAE,OAAO,aAAa,OAAO,SAAS,eAAe,CAAC,MAAM,UAAU,UAAU,MAAM,WAAW,YAAY,QAAQ;AAAA,MACvH;AAAA,IACA,KAAK;AAAY,aAAO;AAAA,QACtB,EAAE,OAAO,cAAc,OAAO,cAAc,eAAe,MAAM,UAAU,QAAQ;AAAA,QACnF,EAAE,OAAO,gBAAgB,OAAO,cAAc,eAAe,OAAO,UAAU,UAAU;AAAA,QACxF,EAAE,OAAO,SAAS,OAAO,IAAI,eAAe,CAAC,MAAM,UAAU,UAAU,OAAO;AAAA,MAChF;AAAA,IACA,KAAK;AAAY,aAAO;AAAA,QACtB,EAAE,OAAO,aAAa,OAAO,uBAAuB,eAAe,MAAM,UAAU,QAAQ;AAAA,QAC3F,EAAE,OAAO,oBAAoB,OAAO,eAAe,eAAe,OAAO,UAAU,UAAU;AAAA,QAC7F,EAAE,OAAO,SAAS,OAAO,IAAI,eAAe,CAAC,MAAM,UAAU,UAAU,OAAO;AAAA,MAChF;AAAA,IACA;AACE,aAAO,kBAAkB,MAAM,WAAW,MAAM,WAAW,MAAM,QAAQ;AAAA,EAC7E;AACF;AAQO,SAAS,SAAS,MAA6D;AACpF,SAAO;AAAA,IACL,WAAW,KAAK;AAAA,IAChB,SAAS,KAAK;AAAA,IACd,YAAY,KAAK,OAAO,IAAI,YAAU,EAAE,OAAO,OAAO,mBAAmB,KAAK,EAAE,EAAE;AAAA,EACpF;AACF;",
+  "names": []
+}

package/dist/interactive-probe.d.ts

@@ -0,0 +1,144 @@
+/**
+ * Interactive Probe — Phase 2 Dynamic Observation
+ *
+ * The static crawl captures DOM structure at rest. This module runs AFTER the static crawl
+ * to observe dynamic behaviour by actually interacting with the UI:
+ *
+ *   1. Empty submit  → captures required-field validation error messages
+ *   2. Invalid format → captures format validation errors (bad email, short password)
+ *   3. Button clicks → captures modal/dialog contents
+ *   4. Select changes → captures conditional reveals (show/hide logic)
+ *   5. Success paths → captures toast/banner messages (where input is safe)
+ *
+ * Results are stored as KnowledgeChunks (sourceType='probe') and update the
+ * ERROR_CATALOGUE + INTERACTION_PATTERNS skill files with real observed data
+ * instead of LLM inference.
+ *
+ * Call signature: probeScreen(page, screenId, elements, opts)
+ * Called from crawlProject() after static element extraction when opts.interactive=true.
+ */
+import type { Page } from 'playwright';
+export interface ProbeElementFact {
+    meaning: string;
+    role: string;
+    expectedData?: string;
+}
+export interface ValidationError {
+    field: string;
+    trigger: 'empty_submit' | 'invalid_format' | 'boundary';
+    errorText: string;
+}
+export interface SuccessMessage {
+    type: 'toast' | 'inline_success' | 'redirect';
+    content: string;
+}
+export interface ModalObservation {
+    triggerElement: string;
+    title: string;
+    content: string;
+    fields: string[];
+    stateId?: string;
+    domSnapshot?: string;
+}
+export interface ConditionalReveal {
+    triggerElement: string;
+    triggerValue: string;
+    revealedText: string;
+    stateId?: string;
+    domSnapshot?: string;
+}
+export interface TooltipCapture {
+    triggerLabel: string;
+    tooltipText: string;
+}
+export interface TabPanelCapture {
+    tabLabel: string;
+    panelContent: string;
+    panelHeading?: string;
+}
+export interface CarouselSlide {
+    slideIndex: number;
+    heading?: string;
+    content: string;
+}
+export interface WizardStep {
+    stepIndex: number;
+    stepLabel?: string;
+    content: string;
+    fields: string[];
+}
+export interface FocusTrapResult {
+    modalTrigger: string;
+    focusTrapped: boolean;
+    sampledFocused: string[];
+}
+export interface ContextMenuCapture {
+    trigger: string;
+    items: string[];
+}
+export interface ProbeResult {
+    screenId: string;
+    url: string;
+    validationErrors: ValidationError[];
+    successMessages: SuccessMessage[];
+    modals: ModalObservation[];
+    conditionals: ConditionalReveal[];
+    loadingPatterns: string[];
+    tooltips?: TooltipCapture[];
+    tabPanels?: TabPanelCapture[];
+    carouselSlides?: CarouselSlide[];
+    wizardSteps?: WizardStep[];
+    focusTraps?: FocusTrapResult[];
+    skipNavIssues?: string[];
+    contextMenus?: ContextMenuCapture[];
+}
+/**
+ * Run interactive probing on a single screen page.
+ * The page must already be navigated to the screen URL.
+ * `elements` is the output of the static extractElements pass.
+ */
+export declare function probeScreen(page: Page, screenId: string, url: string, elements: ProbeElementFact[]): Promise<ProbeResult>;
+/** Phase 5: Convert ProbeResult to typed ProbeObservationFact rows for direct DB storage. */
+export declare function probeResultToObservations(result: ProbeResult): Array<{
+    elementMeaning: string;
+    observationType: string;
+    observedText: string;
+    triggeredBy: string;
+}>;
+/**
+ * Convert a ProbeResult into markdown to store as KnowledgeChunks
+ * and update ERROR_CATALOGUE + INTERACTION_PATTERNS skill files.
+ */
+export declare function probeResultToMarkdown(result: ProbeResult): {
+    errorCatalogueChunk: string;
+    interactionPatternsChunk: string;
+};
+/**
+ * Run interactive probing for all screens discovered in a crawl session.
+ * Called from crawlProject() when opts.interactive = true.
+ */
+export declare function probeAllScreens(page: any, tenantId: string, projectId: string, discovered: Array<{
+    url: string;
+    screenId?: string;
+    elements?: any[];
+}>, onProgress?: (pct: number, step: string) => Promise<void>, jobId?: string): Promise<void>;
+export interface FocusableElement {
+    tag: string;
+    role: string | null;
+    ariaLabel: string | null;
+    text: string | null;
+    hasFocusIndicator: boolean;
+    tabIndex: number | null;
+}
+export interface KeyboardNavResult {
+    focusableCount: number;
+    elementsWithNoFocusIndicator: number;
+    sequence: FocusableElement[];
+    issues: string[];
+}
+/**
+ * Probe keyboard navigation accessibility by pressing Tab up to 50 times and
+ * recording each focused element's properties and focus-indicator visibility.
+ * Stops early if focus wraps back to the start or hits the same element twice.
+ */
+export declare function probeKeyboardNav(page: Page): Promise<KeyboardNavResult>;