@tangle-network/browser-agent-driver 0.30.0 → 0.31.0

package/dist/discover/wayback.js

@@ -0,0 +1,123 @@
+/**
+ * Wayback Machine snapshot discoverer (CDX API).
+ *
+ * Uses archive.org's CDX server to list captures of a URL, then samples
+ * `count` snapshots evenly across the time range. The CDX API is rate-limited
+ * (~15 req/s) but for the small N we use here we never approach the cap.
+ *
+ * CDX response shape:
+ *   [["urlkey","timestamp","original","mimetype","statuscode","digest","length"],
+ *    ["com,stripe)/", "20100101120000", "https://stripe.com/", "text/html", "200", "...", "1234"],
+ *    ...]
+ *
+ * Snapshot URL = `https://web.archive.org/web/<timestamp>/<original>`.
+ */
+const CDX_ENDPOINT = 'https://web.archive.org/cdx/search/cdx';
+const DEFAULT_COUNT = 5;
+/** Convert ISO date "2010-01-01" to CDX yyyymmddhhmmss. */
+function isoToCdxStamp(iso, end = false) {
+    const d = new Date(iso);
+    if (Number.isNaN(d.getTime()))
+        throw new Error(`wayback: invalid date "${iso}"`);
+    const y = d.getUTCFullYear();
+    const m = String(d.getUTCMonth() + 1).padStart(2, '0');
+    const day = String(d.getUTCDate()).padStart(2, '0');
+    return end ? `${y}${m}${day}235959` : `${y}${m}${day}000000`;
+}
+/** Convert CDX timestamp "20100101120000" to ISO datetime. */
+export function cdxStampToIso(stamp) {
+    if (stamp.length < 14)
+        throw new Error(`wayback: malformed CDX timestamp "${stamp}"`);
+    const y = stamp.slice(0, 4);
+    const m = stamp.slice(4, 6);
+    const d = stamp.slice(6, 8);
+    const hh = stamp.slice(8, 10);
+    const mm = stamp.slice(10, 12);
+    const ss = stamp.slice(12, 14);
+    return `${y}-${m}-${d}T${hh}:${mm}:${ss}Z`;
+}
+export function snapshotUrl(timestamp, original) {
+    return `https://web.archive.org/web/${timestamp}/${original}`;
+}
+/** Sample `count` evenly-spaced rows from a sorted list. Always includes first and last. */
+export function sampleEvenly(rows, count) {
+    if (rows.length <= count)
+        return [...rows];
+    if (count <= 0)
+        return [];
+    if (count === 1)
+        return [rows[Math.floor((rows.length - 1) / 2)]];
+    const out = [];
+    const step = (rows.length - 1) / (count - 1);
+    for (let i = 0; i < count; i++) {
+        out.push(rows[Math.round(i * step)]);
+    }
+    return out;
+}
+/** Parse the CDX JSON response (first row is a header). */
+export function parseCdxRows(json) {
+    if (!Array.isArray(json) || json.length < 2)
+        return [];
+    // Skip header row.
+    const rows = json.slice(1);
+    const out = [];
+    for (const r of rows) {
+        if (!Array.isArray(r) || r.length < 7)
+            continue;
+        out.push({
+            urlkey: String(r[0]),
+            timestamp: String(r[1]),
+            original: String(r[2]),
+            mimetype: String(r[3]),
+            statuscode: String(r[4]),
+            digest: String(r[5]),
+            length: String(r[6]),
+        });
+    }
+    return out;
+}
+/** Discover N evenly-spaced wayback snapshots for a URL. */
+export async function discoverWaybackSnapshots(url, opts = {}) {
+    const fetchImpl = opts.fetch ?? globalThis.fetch;
+    if (!fetchImpl)
+        throw new Error('wayback: no fetch implementation available');
+    const count = opts.count ?? DEFAULT_COUNT;
+    const params = new URLSearchParams({
+        url,
+        output: 'json',
+        // collapse=timestamp:6 dedupes to one capture per month (yyyymm = 6 chars).
+        // Without this, CDX returns every capture in the window — which for popular
+        // sites is tens of thousands and silently skews `sampleEvenly` if combined
+        // with a `limit`. With the collapse, the row count is bounded by the
+        // window length in months, so we don't need a limit.
+        collapse: 'timestamp:6',
+    });
+    if (opts.since)
+        params.set('from', isoToCdxStamp(opts.since));
+    if (opts.until)
+        params.set('to', isoToCdxStamp(opts.until, true));
+    if (opts.status200Only !== false) {
+        params.set('filter', 'statuscode:200');
+    }
+    const resp = await fetchImpl(`${CDX_ENDPOINT}?${params.toString()}`);
+    if (!resp.ok) {
+        throw new Error(`wayback: CDX returned ${resp.status} for ${url}`);
+    }
+    const json = await resp.json();
+    const rows = parseCdxRows(json);
+    if (rows.length === 0)
+        return [];
+    // CDX returns chronological order by default. Sample evenly across.
+    const sampled = sampleEvenly(rows, count);
+    return sampled.map(r => ({
+        url,
+        snapshotUrl: snapshotUrl(r.timestamp, r.original),
+        capturedAt: cdxStampToIso(r.timestamp),
+    }));
+}
+/** Expand a list of seed URLs into wayback targets, in parallel. */
+export async function expandWaybackTargets(urls, opts = {}) {
+    const all = await Promise.all(urls.map(u => discoverWaybackSnapshots(u, opts).catch(() => [])));
+    return all.flat();
+}
+//# sourceMappingURL=wayback.js.map

package/dist/discover/wayback.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"wayback.js","sourceRoot":"","sources":["../../src/discover/wayback.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;GAaG;AAIH,MAAM,YAAY,GAAG,wCAAwC,CAAA;AAyB7D,MAAM,aAAa,GAAG,CAAC,CAAA;AAEvB,2DAA2D;AAC3D,SAAS,aAAa,CAAC,GAAW,EAAE,GAAG,GAAG,KAAK;IAC7C,MAAM,CAAC,GAAG,IAAI,IAAI,CAAC,GAAG,CAAC,CAAA;IACvB,IAAI,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,OAAO,EAAE,CAAC;QAAE,MAAM,IAAI,KAAK,CAAC,0BAA0B,GAAG,GAAG,CAAC,CAAA;IAChF,MAAM,CAAC,GAAG,CAAC,CAAC,cAAc,EAAE,CAAA;IAC5B,MAAM,CAAC,GAAG,MAAM,CAAC,CAAC,CAAC,WAAW,EAAE,GAAG,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,EAAE,GAAG,CAAC,CAAA;IACtD,MAAM,GAAG,GAAG,MAAM,CAAC,CAAC,CAAC,UAAU,EAAE,CAAC,CAAC,QAAQ,CAAC,CAAC,EAAE,GAAG,CAAC,CAAA;IACnD,OAAO,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,GAAG,CAAC,GAAG,GAAG,QAAQ,CAAC,CAAC,CAAC,GAAG,CAAC,GAAG,CAAC,GAAG,GAAG,QAAQ,CAAA;AAC9D,CAAC;AAED,8DAA8D;AAC9D,MAAM,UAAU,aAAa,CAAC,KAAa;IACzC,IAAI,KAAK,CAAC,MAAM,GAAG,EAAE;QAAE,MAAM,IAAI,KAAK,CAAC,qCAAqC,KAAK,GAAG,CAAC,CAAA;IACrF,MAAM,CAAC,GAAG,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAA;IAC3B,MAAM,CAAC,GAAG,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAA;IAC3B,MAAM,CAAC,GAAG,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAA;IAC3B,MAAM,EAAE,GAAG,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,CAAA;IAC7B,MAAM,EAAE,GAAG,KAAK,CAAC,KAAK,CAAC,EAAE,EAAE,EAAE,CAAC,CAAA;IAC9B,MAAM,EAAE,GAAG,KAAK,CAAC,KAAK,CAAC,EAAE,EAAE,EAAE,CAAC,CAAA;IAC9B,OAAO,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,GAAG,CAAA;AAC5C,CAAC;AAED,MAAM,UAAU,WAAW,CAAC,SAAiB,EAAE,QAAgB;IAC7D,OAAO,+BAA+B,SAAS,IAAI,QAAQ,EAAE,CAAA;AAC/D,CAAC;AAED,4FAA4F;AAC5F,MAAM,UAAU,YAAY,CAAI,IAAS,EAAE,KAAa;IACtD,IAAI,IAAI,CAAC,MAAM,IAAI,KAAK;QAAE,OAAO,CAAC,GAAG,IAAI,CAAC,CAAA;IAC1C,IAAI,KAAK,IAAI,CAAC;QAAE,OAAO,EAAE,CAAA;IACzB,IAAI,KAAK,KAAK,CAAC;QAAE,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAA;IACjE,MAAM,GAAG,GAAQ,EAAE,CAAA;IACnB,MAAM,IAAI,GAAG,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC,GAAG,CAAC,KAAK,GAAG,CAAC,CAAC,CAAA;IAC5C,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,EAAE,CAAC,EAAE,EAAE,CAAC;QAC/B,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,CAAC,CAAA;IACtC,CAAC;IACD,OAAO,GAAG,CAAA;AACZ,CAAC;AAED,2DAA2D;AAC3D,MAAM,UAAU,YAAY,CAAC,IAAa;IACxC,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,IAAI,CAAC,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC;QAAE,OAAO,EAAE,CAAA;IACtD,mBAAmB;IACnB,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,CAAc,CAAA;IACvC,MAAM,GAAG,GAAa,EAAE,CAAA;IACxB,KAAK,MAAM,CAAC,IAAI,IAAI,EAAE,CAAC;QACrB,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,MAAM,GAAG,CAAC;YAAE,SAAQ;QAC/C,GAAG,CAAC,IAAI,CAAC;YACP,MAAM,EAAE,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;YACpB,SAAS,EAAE,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;YACvB,QAAQ,EAAE,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;YACtB,QAAQ,EAAE,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;YACtB,UAAU,EAAE,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;YACxB,MAAM,EAAE,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;YACpB,MAAM,EAAE,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;SACrB,CAAC,CAAA;IACJ,CAAC;IACD,OAAO,GAAG,CAAA;AACZ,CAAC;AAED,4DAA4D;AAC5D,MAAM,CAAC,KAAK,UAAU,wBAAwB,CAAC,GAAW,EAAE,OAAuB,EAAE;IACnF,MAAM,SAAS,GAAG,IAAI,CAAC,KAAK,IAAI,UAAU,CAAC,KAAK,CAAA;IAChD,IAAI,CAAC,SAAS;QAAE,MAAM,IAAI,KAAK,CAAC,4CAA4C,CAAC,CAAA;IAE7E,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,IAAI,aAAa,CAAA;IACzC,MAAM,MAAM,GAAG,IAAI,eAAe,CAAC;QACjC,GAAG;QACH,MAAM,EAAE,MAAM;QACd,4EAA4E;QAC5E,4EAA4E;QAC5E,2EAA2E;QAC3E,qEAAqE;QACrE,qDAAqD;QACrD,QAAQ,EAAE,aAAa;KACxB,CAAC,CAAA;IACF,IAAI,IAAI,CAAC,KAAK;QAAE,MAAM,CAAC,GAAG,CAAC,MAAM,EAAE,aAAa,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAA;IAC7D,IAAI,IAAI,CAAC,KAAK;QAAE,MAAM,CAAC,GAAG,CAAC,IAAI,EAAE,aAAa,CAAC,IAAI,CAAC,KAAK,EAAE,IAAI,CAAC,CAAC,CAAA;IACjE,IAAI,IAAI,CAAC,aAAa,KAAK,KAAK,EAAE,CAAC;QACjC,MAAM,CAAC,GAAG,CAAC,QAAQ,EAAE,gBAAgB,CAAC,CAAA;IACxC,CAAC;IAED,MAAM,IAAI,GAAG,MAAM,SAAS,CAAC,GAAG,YAAY,IAAI,MAAM,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAA;IACpE,IAAI,CAAC,IAAI,CAAC,EAAE,EAAE,CAAC;QACb,MAAM,IAAI,KAAK,CAAC,yBAAyB,IAAI,CAAC,MAAM,QAAQ,GAAG,EAAE,CAAC,CAAA;IACpE,CAAC;IACD,MAAM,IAAI,GAAG,MAAM,IAAI,CAAC,IAAI,EAAa,CAAA;IACzC,MAAM,IAAI,GAAG,YAAY,CAAC,IAAI,CAAC,CAAA;IAC/B,IAAI,IAAI,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,EAAE,CAAA;IAChC,oEAAoE;IACpE,MAAM,OAAO,GAAG,YAAY,CAAC,IAAI,EAAE,KAAK,CAAC,CAAA;IACzC,OAAO,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;QACvB,GAAG;QACH,WAAW,EAAE,WAAW,CAAC,CAAC,CAAC,SAAS,EAAE,CAAC,CAAC,QAAQ,CAAC;QACjD,UAAU,EAAE,aAAa,CAAC,CAAC,CAAC,SAAS,CAAC;KACvC,CAAC,CAAC,CAAA;AACL,CAAC;AAED,oEAAoE;AACpE,MAAM,CAAC,KAAK,UAAU,oBAAoB,CAAC,IAAc,EAAE,OAAuB,EAAE;IAClF,MAAM,GAAG,GAAG,MAAM,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,wBAAwB,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC,KAAK,CAAC,GAAG,EAAE,CAAC,EAAiB,CAAC,CAAC,CAAC,CAAA;IAC9G,OAAO,GAAG,CAAC,IAAI,EAAE,CAAA;AACnB,CAAC"}

package/dist/jobs/anti-bot.d.ts

@@ -0,0 +1,19 @@
+/**
+ * Anti-bot / blocked-page detection. Pure pattern match against an audit's
+ * report.json — we propagate the existing audit signals rather than re-running
+ * inference.
+ *
+ * Returns a reason string when blocked (so the job can carry it through to
+ * the result envelope), else null.
+ */
+export interface BlockSignals {
+    title?: string;
+    intent?: string;
+    type?: string;
+    ensembleConfidence?: number;
+    findingCount?: number;
+}
+/** Check the audit's report.json for anti-bot patterns. Returns the reason or null. */
+export declare function detectBlock(report: unknown): string | null;
+export declare function reasonFor(s: BlockSignals): string | null;
+//# sourceMappingURL=anti-bot.d.ts.map

package/dist/jobs/anti-bot.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"anti-bot.d.ts","sourceRoot":"","sources":["../../src/jobs/anti-bot.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG;AAsBH,MAAM,WAAW,YAAY;IAC3B,KAAK,CAAC,EAAE,MAAM,CAAA;IACd,MAAM,CAAC,EAAE,MAAM,CAAA;IACf,IAAI,CAAC,EAAE,MAAM,CAAA;IACb,kBAAkB,CAAC,EAAE,MAAM,CAAA;IAC3B,YAAY,CAAC,EAAE,MAAM,CAAA;CACtB;AAED,uFAAuF;AACvF,wBAAgB,WAAW,CAAC,MAAM,EAAE,OAAO,GAAG,MAAM,GAAG,IAAI,CAc1D;AAED,wBAAgB,SAAS,CAAC,CAAC,EAAE,YAAY,GAAG,MAAM,GAAG,IAAI,CAmBxD"}

package/dist/jobs/anti-bot.js

@@ -0,0 +1,64 @@
+/**
+ * Anti-bot / blocked-page detection. Pure pattern match against an audit's
+ * report.json — we propagate the existing audit signals rather than re-running
+ * inference.
+ *
+ * Returns a reason string when blocked (so the job can carry it through to
+ * the result envelope), else null.
+ */
+const TITLE_PATTERNS = [
+    /just a moment\.{3}/i,
+    /^attention required/i,
+    /access denied/i,
+    /verify you are human/i,
+    /enable javascript and cookies/i,
+    /one more step/i,
+    /please complete the security check/i,
+    /^cloudflare/i,
+    /challenge[- ]page/i,
+];
+const INTENT_PATTERNS = [
+    /cloudflare challenge/i,
+    /anti.?bot/i,
+    /captcha/i,
+    /verify (the )?(human|user|browser)/i,
+    /access (denied|restricted|blocked)/i,
+];
+/** Check the audit's report.json for anti-bot patterns. Returns the reason or null. */
+export function detectBlock(report) {
+    const r = report;
+    const page = r.pages?.[0];
+    if (!page)
+        return null;
+    const v2cls = page.auditResultV2?.classification;
+    const cls = v2cls ?? page.classification ?? {};
+    const signals = {
+        title: page.title,
+        intent: cls.intent,
+        type: cls.type,
+        ensembleConfidence: cls.ensembleConfidence,
+        findingCount: page.findings?.length ?? 0,
+    };
+    return reasonFor(signals);
+}
+export function reasonFor(s) {
+    const title = (s.title ?? '').trim();
+    const intent = (s.intent ?? '').trim();
+    if (TITLE_PATTERNS.some(re => re.test(title))) {
+        return `blocked: page title looks like an anti-bot challenge ("${title.slice(0, 80)}")`;
+    }
+    if (INTENT_PATTERNS.some(re => re.test(intent))) {
+        return `blocked: classification intent indicates a challenge page ("${intent.slice(0, 80)}")`;
+    }
+    // Last-resort heuristic: zero findings + very low ensemble confidence + unknown
+    // page-type is overwhelmingly an anti-bot or empty page. Leaving it in the
+    // leaderboard pollutes rankings.
+    if ((s.findingCount ?? 0) === 0
+        && typeof s.ensembleConfidence === 'number'
+        && s.ensembleConfidence < 0.35
+        && s.type === 'unknown') {
+        return 'blocked: zero findings, low classifier confidence, unknown type — likely empty/blocked';
+    }
+    return null;
+}
+//# sourceMappingURL=anti-bot.js.map

package/dist/jobs/anti-bot.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"anti-bot.js","sourceRoot":"","sources":["../../src/jobs/anti-bot.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG;AAEH,MAAM,cAAc,GAAG;IACrB,qBAAqB;IACrB,sBAAsB;IACtB,gBAAgB;IAChB,uBAAuB;IACvB,gCAAgC;IAChC,gBAAgB;IAChB,qCAAqC;IACrC,cAAc;IACd,oBAAoB;CACrB,CAAA;AAED,MAAM,eAAe,GAAG;IACtB,uBAAuB;IACvB,YAAY;IACZ,UAAU;IACV,qCAAqC;IACrC,qCAAqC;CACtC,CAAA;AAUD,uFAAuF;AACvF,MAAM,UAAU,WAAW,CAAC,MAAe;IACzC,MAAM,CAAC,GAAG,MAA6P,CAAA;IACvQ,MAAM,IAAI,GAAG,CAAC,CAAC,KAAK,EAAE,CAAC,CAAC,CAAC,CAAA;IACzB,IAAI,CAAC,IAAI;QAAE,OAAO,IAAI,CAAA;IACtB,MAAM,KAAK,GAAG,IAAI,CAAC,aAAa,EAAE,cAAc,CAAA;IAChD,MAAM,GAAG,GAAG,KAAK,IAAI,IAAI,CAAC,cAAc,IAAI,EAAE,CAAA;IAC9C,MAAM,OAAO,GAAiB;QAC5B,KAAK,EAAE,IAAI,CAAC,KAAK;QACjB,MAAM,EAAE,GAAG,CAAC,MAAM;QAClB,IAAI,EAAE,GAAG,CAAC,IAAI;QACd,kBAAkB,EAAE,GAAG,CAAC,kBAAkB;QAC1C,YAAY,EAAE,IAAI,CAAC,QAAQ,EAAE,MAAM,IAAI,CAAC;KACzC,CAAA;IACD,OAAO,SAAS,CAAC,OAAO,CAAC,CAAA;AAC3B,CAAC;AAED,MAAM,UAAU,SAAS,CAAC,CAAe;IACvC,MAAM,KAAK,GAAG,CAAC,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAA;IACpC,MAAM,MAAM,GAAG,CAAC,CAAC,CAAC,MAAM,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAA;IACtC,IAAI,cAAc,CAAC,IAAI,CAAC,EAAE,CAAC,EAAE,CAAC,EAAE,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC;QAC9C,OAAO,0DAA0D,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,IAAI,CAAA;IACzF,CAAC;IACD,IAAI,eAAe,CAAC,IAAI,CAAC,EAAE,CAAC,EAAE,CAAC,EAAE,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,EAAE,CAAC;QAChD,OAAO,+DAA+D,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,IAAI,CAAA;IAC/F,CAAC;IACD,gFAAgF;IAChF,2EAA2E;IAC3E,iCAAiC;IACjC,IAAI,CAAC,CAAC,CAAC,YAAY,IAAI,CAAC,CAAC,KAAK,CAAC;WAC1B,OAAO,CAAC,CAAC,kBAAkB,KAAK,QAAQ;WACxC,CAAC,CAAC,kBAAkB,GAAG,IAAI;WAC3B,CAAC,CAAC,IAAI,KAAK,SAAS,EAAE,CAAC;QAC1B,OAAO,wFAAwF,CAAA;IACjG,CAAC;IACD,OAAO,IAAI,CAAA;AACb,CAAC"}

package/dist/jobs/cost-estimate.d.ts

@@ -0,0 +1,12 @@
+/**
+ * Pre-flight cost estimate — refuses to silently spend $1k.
+ *
+ * The numbers here are deliberately conservative and based on observed
+ * production telemetry for `claude-code` provider with sonnet at the time of
+ * writing. If the assumptions drift, only one constant changes.
+ */
+import type { CostEstimate, JobSpec } from './types.js';
+/** Average USD per single-page audit. Tuned for sonnet via claude-code provider, 1 audit pass. */
+export declare const DEFAULT_PER_AUDIT_USD = 0.4;
+export declare function estimateCost(spec: JobSpec, targetCount: number, perAuditUSD?: number): CostEstimate;
+//# sourceMappingURL=cost-estimate.d.ts.map

package/dist/jobs/cost-estimate.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"cost-estimate.d.ts","sourceRoot":"","sources":["../../src/jobs/cost-estimate.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AAEH,OAAO,KAAK,EAAE,YAAY,EAAE,OAAO,EAAE,MAAM,YAAY,CAAA;AAEvD,kGAAkG;AAClG,eAAO,MAAM,qBAAqB,MAAM,CAAA;AAExC,wBAAgB,YAAY,CAAC,IAAI,EAAE,OAAO,EAAE,WAAW,EAAE,MAAM,EAAE,WAAW,SAAwB,GAAG,YAAY,CAUlH"}

package/dist/jobs/cost-estimate.js

@@ -0,0 +1,21 @@
+/**
+ * Pre-flight cost estimate — refuses to silently spend $1k.
+ *
+ * The numbers here are deliberately conservative and based on observed
+ * production telemetry for `claude-code` provider with sonnet at the time of
+ * writing. If the assumptions drift, only one constant changes.
+ */
+/** Average USD per single-page audit. Tuned for sonnet via claude-code provider, 1 audit pass. */
+export const DEFAULT_PER_AUDIT_USD = 0.4;
+export function estimateCost(spec, targetCount, perAuditUSD = DEFAULT_PER_AUDIT_USD) {
+    const pages = Math.max(spec.audit?.pages ?? 1, 1);
+    const estimatedTotalUSD = targetCount * pages * perAuditUSD;
+    const exceedsCap = typeof spec.maxCostUSD === 'number' && estimatedTotalUSD > spec.maxCostUSD;
+    return {
+        targetCount,
+        perAuditUSD: perAuditUSD * pages,
+        estimatedTotalUSD,
+        exceedsCap,
+    };
+}
+//# sourceMappingURL=cost-estimate.js.map

package/dist/jobs/cost-estimate.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"cost-estimate.js","sourceRoot":"","sources":["../../src/jobs/cost-estimate.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AAIH,kGAAkG;AAClG,MAAM,CAAC,MAAM,qBAAqB,GAAG,GAAG,CAAA;AAExC,MAAM,UAAU,YAAY,CAAC,IAAa,EAAE,WAAmB,EAAE,WAAW,GAAG,qBAAqB;IAClG,MAAM,KAAK,GAAG,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,KAAK,EAAE,KAAK,IAAI,CAAC,EAAE,CAAC,CAAC,CAAA;IACjD,MAAM,iBAAiB,GAAG,WAAW,GAAG,KAAK,GAAG,WAAW,CAAA;IAC3D,MAAM,UAAU,GAAG,OAAO,IAAI,CAAC,UAAU,KAAK,QAAQ,IAAI,iBAAiB,GAAG,IAAI,CAAC,UAAU,CAAA;IAC7F,OAAO;QACL,WAAW;QACX,WAAW,EAAE,WAAW,GAAG,KAAK;QAChC,iBAAiB;QACjB,UAAU;KACX,CAAA;AACH,CAAC"}

package/dist/jobs/cost-history.d.ts

@@ -0,0 +1,19 @@
+/**
+ * Adaptive cost estimate from historical jobs. The default flat
+ * `DEFAULT_PER_AUDIT_USD` is still a fine starting point for a fresh user,
+ * but once 3+ jobs have completed we can do better: averaging the actual
+ * per-target cost across recent jobs is closer to ground truth, especially
+ * once ethics / first-principles modes start firing differently per target.
+ *
+ * Pure function of `~/.bad/jobs/` records — no telemetry endpoint required.
+ */
+export interface AdaptiveCostStats {
+    perAuditUSD: number;
+    source: 'history' | 'default';
+    /** Number of historical job records the estimate was averaged over. */
+    jobsObserved: number;
+    /** Number of audited targets the estimate was averaged over. */
+    targetsObserved: number;
+}
+export declare function computePerAuditFromHistory(dir?: string): AdaptiveCostStats;
+//# sourceMappingURL=cost-history.d.ts.map

package/dist/jobs/cost-history.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"cost-history.d.ts","sourceRoot":"","sources":["../../src/jobs/cost-history.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AASH,MAAM,WAAW,iBAAiB;IAChC,WAAW,EAAE,MAAM,CAAA;IACnB,MAAM,EAAE,SAAS,GAAG,SAAS,CAAA;IAC7B,uEAAuE;IACvE,YAAY,EAAE,MAAM,CAAA;IACpB,gEAAgE;IAChE,eAAe,EAAE,MAAM,CAAA;CACxB;AAED,wBAAgB,0BAA0B,CAAC,GAAG,CAAC,EAAE,MAAM,GAAG,iBAAiB,CAyB1E"}

package/dist/jobs/cost-history.js

@@ -0,0 +1,42 @@
+/**
+ * Adaptive cost estimate from historical jobs. The default flat
+ * `DEFAULT_PER_AUDIT_USD` is still a fine starting point for a fresh user,
+ * but once 3+ jobs have completed we can do better: averaging the actual
+ * per-target cost across recent jobs is closer to ground truth, especially
+ * once ethics / first-principles modes start firing differently per target.
+ *
+ * Pure function of `~/.bad/jobs/` records — no telemetry endpoint required.
+ */
+import { listJobs, loadJob } from './store.js';
+import { DEFAULT_PER_AUDIT_USD } from './cost-estimate.js';
+/** Min number of completed jobs before we trust history over the static default. */
+const MIN_HISTORY = 3;
+export function computePerAuditFromHistory(dir) {
+    const entries = listJobs(dir);
+    // Only count completed/partial jobs — failed ones have skewed cost.
+    const usable = entries.filter((e) => e.status === 'completed' || e.status === 'partial').slice(0, 20);
+    if (usable.length < MIN_HISTORY) {
+        return { perAuditUSD: DEFAULT_PER_AUDIT_USD, source: 'default', jobsObserved: usable.length, targetsObserved: 0 };
+    }
+    let totalCost = 0;
+    let totalTargets = 0;
+    for (const entry of usable) {
+        const job = loadJob(entry.jobId, dir);
+        if (!job)
+            continue;
+        const okCount = job.results.filter(r => r.status === 'ok' && typeof r.costUSD === 'number').length;
+        if (okCount === 0)
+            continue;
+        totalCost += job.totalCostUSD;
+        totalTargets += okCount;
+    }
+    if (totalTargets === 0) {
+        return { perAuditUSD: DEFAULT_PER_AUDIT_USD, source: 'default', jobsObserved: usable.length, targetsObserved: 0 };
+    }
+    const perAudit = totalCost / totalTargets;
+    // Floor at half the static default to prevent runaway optimism on a stretch
+    // of zero-cost jobs (which can happen with the claude-code provider).
+    const floored = Math.max(perAudit, DEFAULT_PER_AUDIT_USD * 0.5);
+    return { perAuditUSD: floored, source: 'history', jobsObserved: usable.length, targetsObserved: totalTargets };
+}
+//# sourceMappingURL=cost-history.js.map

package/dist/jobs/cost-history.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"cost-history.js","sourceRoot":"","sources":["../../src/jobs/cost-history.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AAGH,OAAO,EAAE,QAAQ,EAAE,OAAO,EAAE,MAAM,YAAY,CAAA;AAC9C,OAAO,EAAE,qBAAqB,EAAE,MAAM,oBAAoB,CAAA;AAE1D,oFAAoF;AACpF,MAAM,WAAW,GAAG,CAAC,CAAA;AAWrB,MAAM,UAAU,0BAA0B,CAAC,GAAY;IACrD,MAAM,OAAO,GAAG,QAAQ,CAAC,GAAG,CAAC,CAAA;IAC7B,oEAAoE;IACpE,MAAM,MAAM,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC,CAAgB,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,KAAK,WAAW,IAAI,CAAC,CAAC,MAAM,KAAK,SAAS,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,CAAA;IACpH,IAAI,MAAM,CAAC,MAAM,GAAG,WAAW,EAAE,CAAC;QAChC,OAAO,EAAE,WAAW,EAAE,qBAAqB,EAAE,MAAM,EAAE,SAAS,EAAE,YAAY,EAAE,MAAM,CAAC,MAAM,EAAE,eAAe,EAAE,CAAC,EAAE,CAAA;IACnH,CAAC;IACD,IAAI,SAAS,GAAG,CAAC,CAAA;IACjB,IAAI,YAAY,GAAG,CAAC,CAAA;IACpB,KAAK,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;QAC3B,MAAM,GAAG,GAAG,OAAO,CAAC,KAAK,CAAC,KAAK,EAAE,GAAG,CAAC,CAAA;QACrC,IAAI,CAAC,GAAG;YAAE,SAAQ;QAClB,MAAM,OAAO,GAAG,GAAG,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,MAAM,KAAK,IAAI,IAAI,OAAO,CAAC,CAAC,OAAO,KAAK,QAAQ,CAAC,CAAC,MAAM,CAAA;QAClG,IAAI,OAAO,KAAK,CAAC;YAAE,SAAQ;QAC3B,SAAS,IAAI,GAAG,CAAC,YAAY,CAAA;QAC7B,YAAY,IAAI,OAAO,CAAA;IACzB,CAAC;IACD,IAAI,YAAY,KAAK,CAAC,EAAE,CAAC;QACvB,OAAO,EAAE,WAAW,EAAE,qBAAqB,EAAE,MAAM,EAAE,SAAS,EAAE,YAAY,EAAE,MAAM,CAAC,MAAM,EAAE,eAAe,EAAE,CAAC,EAAE,CAAA;IACnH,CAAC;IACD,MAAM,QAAQ,GAAG,SAAS,GAAG,YAAY,CAAA;IACzC,4EAA4E;IAC5E,sEAAsE;IACtE,MAAM,OAAO,GAAG,IAAI,CAAC,GAAG,CAAC,QAAQ,EAAE,qBAAqB,GAAG,GAAG,CAAC,CAAA;IAC/D,OAAO,EAAE,WAAW,EAAE,OAAO,EAAE,MAAM,EAAE,SAAS,EAAE,YAAY,EAAE,MAAM,CAAC,MAAM,EAAE,eAAe,EAAE,YAAY,EAAE,CAAA;AAChH,CAAC"}

package/dist/jobs/index.d.ts

@@ -0,0 +1,31 @@
+/**
+ * Jobs — declarative comparative-audit jobs.
+ *
+ * Public surface:
+ *   createJob(spec) → Job (queued, persisted, ready to run)
+ *   runJob(job, opts) → Job (executes the fan-out, returns final state)
+ *   loadJob / listJobs / saveJob — store accessors
+ *   estimateCost — pre-flight cost guard
+ */
+export type { Job, JobSpec, JobStatus, JobKind, JobTarget, JobResultEntry, JobResultStatus, DiscoverSpec, DiscoverSource, AuditOptions, CostEstimate, } from './types.js';
+export { newJobId, saveJob, loadJob, listJobs, appendIndexEntry, jobsDir, jobPath, updateJobStatus } from './store.js';
+export type { JobIndexEntry } from './store.js';
+export { estimateCost, DEFAULT_PER_AUDIT_USD } from './cost-estimate.js';
+export { runJob } from './queue.js';
+export type { AuditFn, RunJobOptions } from './queue.js';
+export { withRetry, isRetryableDefault, DEFAULT_RETRY_POLICY } from './retry.js';
+export type { RetryPolicy } from './retry.js';
+export { detectBlock, reasonFor } from './anti-bot.js';
+export type { BlockSignals } from './anti-bot.js';
+export { computePerAuditFromHistory } from './cost-history.js';
+export type { AdaptiveCostStats } from './cost-history.js';
+export { orchestrateJob, needsIntervention } from './orchestrator.js';
+export type { OrchestrateJobOptions } from './orchestrator.js';
+import type { Job, JobSpec, JobTarget } from './types.js';
+/**
+ * Mint a fresh `queued` job from a spec. Targets are seeded from the spec
+ * (snapshot expansion, if any, must be done by the caller via the
+ * `discover` module before runJob).
+ */
+export declare function createJob(spec: JobSpec, targets: JobTarget[], dir?: string): Job;
+//# sourceMappingURL=index.d.ts.map

package/dist/jobs/index.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../src/jobs/index.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AAEH,YAAY,EACV,GAAG,EACH,OAAO,EACP,SAAS,EACT,OAAO,EACP,SAAS,EACT,cAAc,EACd,eAAe,EACf,YAAY,EACZ,cAAc,EACd,YAAY,EACZ,YAAY,GACb,MAAM,YAAY,CAAA;AAEnB,OAAO,EAAE,QAAQ,EAAE,OAAO,EAAE,OAAO,EAAE,QAAQ,EAAE,gBAAgB,EAAE,OAAO,EAAE,OAAO,EAAE,eAAe,EAAE,MAAM,YAAY,CAAA;AACtH,YAAY,EAAE,aAAa,EAAE,MAAM,YAAY,CAAA;AAC/C,OAAO,EAAE,YAAY,EAAE,qBAAqB,EAAE,MAAM,oBAAoB,CAAA;AACxE,OAAO,EAAE,MAAM,EAAE,MAAM,YAAY,CAAA;AACnC,YAAY,EAAE,OAAO,EAAE,aAAa,EAAE,MAAM,YAAY,CAAA;AACxD,OAAO,EAAE,SAAS,EAAE,kBAAkB,EAAE,oBAAoB,EAAE,MAAM,YAAY,CAAA;AAChF,YAAY,EAAE,WAAW,EAAE,MAAM,YAAY,CAAA;AAC7C,OAAO,EAAE,WAAW,EAAE,SAAS,EAAE,MAAM,eAAe,CAAA;AACtD,YAAY,EAAE,YAAY,EAAE,MAAM,eAAe,CAAA;AACjD,OAAO,EAAE,0BAA0B,EAAE,MAAM,mBAAmB,CAAA;AAC9D,YAAY,EAAE,iBAAiB,EAAE,MAAM,mBAAmB,CAAA;AAC1D,OAAO,EAAE,cAAc,EAAE,iBAAiB,EAAE,MAAM,mBAAmB,CAAA;AACrE,YAAY,EAAE,qBAAqB,EAAE,MAAM,mBAAmB,CAAA;AAE9D,OAAO,KAAK,EAAE,GAAG,EAAE,OAAO,EAAE,SAAS,EAAE,MAAM,YAAY,CAAA;AAGzD;;;;GAIG;AACH,wBAAgB,SAAS,CAAC,IAAI,EAAE,OAAO,EAAE,OAAO,EAAE,SAAS,EAAE,EAAE,GAAG,CAAC,EAAE,MAAM,GAAG,GAAG,CAahF"}

package/dist/jobs/index.js

@@ -0,0 +1,37 @@
+/**
+ * Jobs — declarative comparative-audit jobs.
+ *
+ * Public surface:
+ *   createJob(spec) → Job (queued, persisted, ready to run)
+ *   runJob(job, opts) → Job (executes the fan-out, returns final state)
+ *   loadJob / listJobs / saveJob — store accessors
+ *   estimateCost — pre-flight cost guard
+ */
+export { newJobId, saveJob, loadJob, listJobs, appendIndexEntry, jobsDir, jobPath, updateJobStatus } from './store.js';
+export { estimateCost, DEFAULT_PER_AUDIT_USD } from './cost-estimate.js';
+export { runJob } from './queue.js';
+export { withRetry, isRetryableDefault, DEFAULT_RETRY_POLICY } from './retry.js';
+export { detectBlock, reasonFor } from './anti-bot.js';
+export { computePerAuditFromHistory } from './cost-history.js';
+export { orchestrateJob, needsIntervention } from './orchestrator.js';
+import { newJobId, saveJob, appendIndexEntry } from './store.js';
+/**
+ * Mint a fresh `queued` job from a spec. Targets are seeded from the spec
+ * (snapshot expansion, if any, must be done by the caller via the
+ * `discover` module before runJob).
+ */
+export function createJob(spec, targets, dir) {
+    const job = {
+        jobId: newJobId(),
+        spec,
+        status: 'queued',
+        createdAt: new Date().toISOString(),
+        targets,
+        results: [],
+        totalCostUSD: 0,
+    };
+    saveJob(job, dir);
+    appendIndexEntry(job, dir);
+    return job;
+}
+//# sourceMappingURL=index.js.map

package/dist/jobs/index.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.js","sourceRoot":"","sources":["../../src/jobs/index.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AAgBH,OAAO,EAAE,QAAQ,EAAE,OAAO,EAAE,OAAO,EAAE,QAAQ,EAAE,gBAAgB,EAAE,OAAO,EAAE,OAAO,EAAE,eAAe,EAAE,MAAM,YAAY,CAAA;AAEtH,OAAO,EAAE,YAAY,EAAE,qBAAqB,EAAE,MAAM,oBAAoB,CAAA;AACxE,OAAO,EAAE,MAAM,EAAE,MAAM,YAAY,CAAA;AAEnC,OAAO,EAAE,SAAS,EAAE,kBAAkB,EAAE,oBAAoB,EAAE,MAAM,YAAY,CAAA;AAEhF,OAAO,EAAE,WAAW,EAAE,SAAS,EAAE,MAAM,eAAe,CAAA;AAEtD,OAAO,EAAE,0BAA0B,EAAE,MAAM,mBAAmB,CAAA;AAE9D,OAAO,EAAE,cAAc,EAAE,iBAAiB,EAAE,MAAM,mBAAmB,CAAA;AAIrE,OAAO,EAAE,QAAQ,EAAE,OAAO,EAAE,gBAAgB,EAAE,MAAM,YAAY,CAAA;AAEhE;;;;GAIG;AACH,MAAM,UAAU,SAAS,CAAC,IAAa,EAAE,OAAoB,EAAE,GAAY;IACzE,MAAM,GAAG,GAAQ;QACf,KAAK,EAAE,QAAQ,EAAE;QACjB,IAAI;QACJ,MAAM,EAAE,QAAQ;QAChB,SAAS,EAAE,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE;QACnC,OAAO;QACP,OAAO,EAAE,EAAE;QACX,YAAY,EAAE,CAAC;KAChB,CAAA;IACD,OAAO,CAAC,GAAG,EAAE,GAAG,CAAC,CAAA;IACjB,gBAAgB,CAAC,GAAG,EAAE,GAAG,CAAC,CAAA;IAC1B,OAAO,GAAG,CAAA;AACZ,CAAC"}

package/dist/jobs/orchestrator.d.ts

@@ -0,0 +1,42 @@
+/**
+ * Agentic orchestrator — wraps `runJob` with a control loop that handles
+ * strategic decisions (re-sample broken wayback snapshots, retry vs skip,
+ * widen the window when too many fail, conclude) via LLM tool calls.
+ *
+ * Design: protocols are deterministic (retry/backoff lives in `retry.ts`,
+ * anti-bot detection lives in `anti-bot.ts`). The orchestrator only steps
+ * in for *judgment* calls where a hand-tuned heuristic would break the
+ * moment the world changes. Cost ceiling: hard-capped at spec.maxCostUSD,
+ * and each control-loop iteration is a single LLM call (~1.5s, ~$0.005).
+ *
+ * Tool surface (LLM-callable):
+ *   getJobState()        — what's done, what failed, current cost
+ *   resampleWayback(...) — replace targets for a URL with a new sample
+ *   retryTarget(...)     — re-attempt a single failed/skipped target
+ *   markSkipped(...)     — terminal skip with reason
+ *   concludeJob()        — exit the loop
+ *
+ * The orchestrator runs the initial fan-out via `runJob` first (so the
+ * deterministic happy path doesn't pay LLM tax), then enters the loop only
+ * if the result needs intervention.
+ */
+import { Brain } from '../brain/index.js';
+import { type AuditFn } from './queue.js';
+import type { Job } from './types.js';
+export interface OrchestrateJobOptions {
+    auditFn: AuditFn;
+    /** Persistence dir override (tests). */
+    dir?: string;
+    /** Cap on agent control-loop iterations. Default 8. */
+    maxIterations?: number;
+    /** Override the LLM provider (tests). Defaults to the default Brain. */
+    brain?: Brain;
+    /** When true, log every tool call. */
+    verbose?: boolean;
+}
+export declare function orchestrateJob(job: Job, opts: OrchestrateJobOptions): Promise<Job>;
+/** Cheap deterministic check: did the fan-out leave anything worth fixing? */
+declare function needsIntervention(job: Job): boolean;
+/** Re-export for tests / callers that want to inspect what would have run. */
+export { needsIntervention };
+//# sourceMappingURL=orchestrator.d.ts.map

package/dist/jobs/orchestrator.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"orchestrator.d.ts","sourceRoot":"","sources":["../../src/jobs/orchestrator.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;GAqBG;AAGH,OAAO,EAAE,KAAK,EAAE,MAAM,mBAAmB,CAAA;AAEzC,OAAO,EAAU,KAAK,OAAO,EAAE,MAAM,YAAY,CAAA;AAGjD,OAAO,KAAK,EAAE,GAAG,EAA6B,MAAM,YAAY,CAAA;AAEhE,MAAM,WAAW,qBAAqB;IACpC,OAAO,EAAE,OAAO,CAAA;IAChB,wCAAwC;IACxC,GAAG,CAAC,EAAE,MAAM,CAAA;IACZ,uDAAuD;IACvD,aAAa,CAAC,EAAE,MAAM,CAAA;IACtB,wEAAwE;IACxE,KAAK,CAAC,EAAE,KAAK,CAAA;IACb,sCAAsC;IACtC,OAAO,CAAC,EAAE,OAAO,CAAA;CAClB;AAkCD,wBAAsB,cAAc,CAAC,GAAG,EAAE,GAAG,EAAE,IAAI,EAAE,qBAAqB,GAAG,OAAO,CAAC,GAAG,CAAC,CAmDxF;AAED,8EAA8E;AAC9E,iBAAS,iBAAiB,CAAC,GAAG,EAAE,GAAG,GAAG,OAAO,CAa5C;AA+JD,8EAA8E;AAC9E,OAAO,EAAE,iBAAiB,EAAE,CAAA"}