@obsrviq/tracker 0.4.2 → 0.5.0

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@obsrviq/tracker",
-  "version": "0.4.2",
+  "version": "0.5.0",
   "license": "MIT",
   "publishConfig": {
     "access": "public"
@@ -27,7 +27,7 @@
     "@rrweb/types": "2.0.1",
     "fflate": "^0.8.2",
     "rrweb": "2.0.1",
-    "@obsrviq/types": "0.4.0"
+    "@obsrviq/types": "0.4.1"
   },
   "devDependencies": {
     "tsup": "^8.3.5",

package/src/index.ts

@@ -4,6 +4,8 @@ import { Transport } from './transport.js';
 import { instrumentDom } from './recorder.js';
 import { instrumentConsole } from './console.js';
 import { instrumentNetwork } from './network.js';
+import { instrumentWebSocket } from './websocket.js';
+import { setCustomMaskKeys, markMaskConfigReady } from './mask.js';
 import { instrumentErrors } from './errors.js';
 import { instrumentVitals } from './vitals.js';
 import { instrumentInteractions } from './interactions.js';
@@ -97,6 +99,14 @@ class ObsrviqClient {
     // askPermission = "show a built-in consent prompt"; it implies a consent gate.
     if (this.config.askPermission) this.config.requireConsent = true;
 
+    // Custom masking keys: apply any static ones NOW, then fetch the per-site list
+    // (GET /v1/config) and merge. Fired before the network instrumenter is installed
+    // so this request is never self-recorded; the transport also re-masks custom
+    // keys at send time so nothing captured before the fetch resolves can leak.
+    const staticMaskKeys = Array.isArray(config.maskKeys) ? config.maskKeys : [];
+    setCustomMaskKeys(staticMaskKeys);
+    void this.loadRemoteMaskKeys(staticMaskKeys);
+
     this.startSession();
 
     // Init-time identity (rarely known this early — usually set later via identify()).
@@ -124,6 +134,39 @@ class ObsrviqClient {
     }
   }
 
+  /** Fetch the per-site masking config (GET /v1/config) and merge its keys with any
+   *  static ones. Best-effort: on failure the built-in masking still applies. */
+  private async loadRemoteMaskKeys(staticKeys: string[]): Promise<void> {
+    const cfg = this.config;
+    if (!cfg || typeof fetch === 'undefined') {
+      markMaskConfigReady();
+      return;
+    }
+    try {
+      const ctrl = new AbortController();
+      const timer = setTimeout(() => ctrl.abort(), 2000);
+      try {
+        const url = `${cfg.endpoint.replace(/\/$/, '')}/v1/config?key=${encodeURIComponent(cfg.siteKey)}`;
+        const res = await fetch(url, { credentials: 'omit', mode: 'cors', signal: ctrl.signal });
+        if (res.ok) {
+          const data = (await res.json()) as { maskKeys?: unknown };
+          const remote = Array.isArray(data.maskKeys)
+            ? data.maskKeys.filter((k): k is string => typeof k === 'string')
+            : [];
+          setCustomMaskKeys([...staticKeys, ...remote]);
+        }
+      } finally {
+        clearTimeout(timer);
+      }
+    } catch {
+      /* offline / blocked / timeout / bad JSON — built-in masking still protects. */
+    } finally {
+      // Always release the transport's hold — a failed/slow fetch must never block
+      // sending forever (built-in + static-key masking still applies).
+      markMaskConfigReady();
+    }
+  }
+
   /** (Re)create the session id, clock, meta, and transport. Used by init() and
    *  reset(). Does not start recording — callers decide via maybeStart().
    *  @param seedInitUser seed identity from the init-time `cfg.userId`. True for
@@ -237,6 +280,7 @@ class ObsrviqClient {
     // before app code runs more requests; DOM last so its snapshot is current.
     this.teardowns.push(instrumentConsole(ctx));
     this.teardowns.push(instrumentNetwork(ctx));
+    this.teardowns.push(instrumentWebSocket(ctx));
     this.teardowns.push(instrumentErrors(ctx));
     this.teardowns.push(instrumentVitals(ctx));
     this.teardowns.push(instrumentInteractions(ctx));

package/src/mask.ts

@@ -2,14 +2,34 @@
 // anything leaves the page. Heuristics catch the obvious PII; tenants tighten
 // with maskTextSelector / beforeSend.
 
+import type { ObsrviqEvent } from '@obsrviq/types';
+
 const EMAIL = /[a-z0-9._%+-]+@[a-z0-9.-]+\.[a-z]{2,}/gi;
 const CC = /\b(?:\d[ -]*?){13,19}\b/g; // candidate card numbers
 const SSN = /\b\d{3}-\d{2}-\d{4}\b/g;
 const LONG_DIGITS = /\b\d{9,}\b/g; // phone-ish / account-ish
 
-// Keys whose values we always redact regardless of content.
-const SENSITIVE_KEYS =
-  /(pass(word)?|secret|token|api[-_]?key|authorization|cookie|ssn|card|cvv|pin|otp)/i;
+// Keys whose values we always redact, regardless of content. Two tiers so we
+// catch real secrets without over-redacting innocent fields:
+//   STRONG — long/unambiguous tokens, safe to match as a substring of the key
+//            name (e.g. "accessToken", "x-api-key", "clientSecret").
+//   WEAK   — short/ambiguous tokens that we only honour when they stand alone as
+//            a key SEGMENT, so "shipping"(pin) / "discard"(card) / "author"(auth)
+//            aren't caught but "pin" / "cardNumber" / "authToken" are.
+const SENSITIVE_STRONG =
+  /(password|passwd|passphrase|secret|token|authoriz|apikey|api[_-]key|creditcard|cardnumber|cardholder|cookie|cvv|cvc)/i;
+const SENSITIVE_WEAK = new Set(['pass', 'pwd', 'auth', 'card', 'pin', 'otp', 'ssn', 'cvv', 'cvc']);
+
+/** Split a key name into lowercase word segments on delimiters AND camelCase
+ *  boundaries: "accessToken" → [access, token], "api_key" → [api, key]. */
+function keySegments(key: string): string[] {
+  return key
+    .replace(/([a-z0-9])([A-Z])/g, '$1 $2')
+    .replace(/([A-Z]+)([A-Z][a-z])/g, '$1 $2')
+    .split(/[^A-Za-z0-9]+/)
+    .filter(Boolean)
+    .map((s) => s.toLowerCase());
+}
 
 export function maskString(input: string): string {
   return input
@@ -24,8 +44,109 @@ function luhnish(s: string): boolean {
   return digits.length >= 13 && digits.length <= 19;
 }
 
-export function isSensitiveKey(key: string): boolean {
-  return SENSITIVE_KEYS.test(key);
+// Customer-defined keys to mask (case-insensitive exact match), delivered via the
+// per-site config (GET /v1/config) and/or the init `maskKeys` option. Additive to
+// the built-in list above — you can only add protections, never remove them.
+let customKeys = new Set<string>();
+export function setCustomMaskKeys(keys: string[]): void {
+  customKeys = new Set(keys.map((k) => String(k).trim().toLowerCase()).filter(Boolean));
+}
+
+// Whether the per-site mask config (GET /v1/config) has settled. Until it does,
+// remote-only custom keys aren't known yet, so the transport holds non-unload
+// sends (below) to avoid leaking a custom-keyed value in the brief init window.
+let maskConfigSettled = false;
+export function markMaskConfigReady(): void {
+  maskConfigSettled = true;
+}
+export function isMaskConfigPending(): boolean {
+  return !maskConfigSettled;
+}
+export function hasCustomMaskKeys(): boolean {
+  return customKeys.size > 0;
+}
+function isCustomKey(key: string): boolean {
+  return customKeys.has(key.toLowerCase());
+}
+
+/** Key-based sensitivity test: built-in STRONG/WEAK tiers PLUS any customer-
+ *  configured keys (exact, case-insensitive). Boundary-aware so innocent fields
+ *  like "shipping" / "discard" / "author" aren't redacted. */
+export function isSensitiveKeyName(key: string): boolean {
+  if (isCustomKey(key)) return true;
+  if (SENSITIVE_STRONG.test(key)) return true;
+  return keySegments(key).some((seg) => SENSITIVE_WEAK.has(seg));
+}
+
+/** Redact sensitive-key values in a parsed object (built-in list + custom keys).
+ *  The send-time safety net that guarantees a `{"password": "…"}` style body can
+ *  never leave the page unmasked, even for non-DOM network/WS payloads (whose
+ *  capture-time masking is PII-pattern only, not key-aware). */
+function redactSensitiveInValue(value: unknown, depth = 0): unknown {
+  if (depth > 24 || value === null || typeof value !== 'object') return value;
+  if (Array.isArray(value)) return value.map((v) => redactSensitiveInValue(v, depth + 1));
+  const out: Record<string, unknown> = {};
+  for (const [k, v] of Object.entries(value as Record<string, unknown>)) {
+    out[k] = isSensitiveKeyName(k) ? '«redacted»' : redactSensitiveInValue(v, depth + 1);
+  }
+  return out;
+}
+
+/** String-level key redaction for payloads that don't parse as JSON — non-JSON
+ *  bodies (form-urlencoded / query strings) AND truncated JSON, where the
+ *  structured path bails. Captures each `"key": value` (JSON-ish, tolerant of a
+ *  cut-off tail) and `key=value` (form / query) key name and redacts the value
+ *  when the key is sensitive. */
+function redactSensitiveInRawText(text: string): string {
+  let out = text.replace(
+    /"([A-Za-z0-9_.\-]+)"\s*:\s*("(?:[^"\\]|\\.)*"|[^,}\]\s]+)/g,
+    (m, key: string) => (isSensitiveKeyName(key) ? '"' + key + '": "«redacted»"' : m),
+  );
+  out = out.replace(
+    /(^|[?&])([A-Za-z0-9_.\-[\]]+)=([^&\s]*)/g,
+    (m, pre: string, key: string) => (isSensitiveKeyName(key) ? pre + key + '=«redacted»' : m),
+  );
+  return out;
+}
+
+/** Redact sensitive-key values in a stored body/arg string. Tries structured
+ *  JSON first; falls back to a string scan for non-JSON or truncated payloads so a
+ *  sensitive key can't slip through in `email=a&password=b` or a cut-off body. */
+function redactSensitiveInText(text: string | undefined): string | undefined {
+  if (!text) return text;
+  const trimmed = text.trim();
+  if (trimmed[0] === '{' || trimmed[0] === '[') {
+    try {
+      return JSON.stringify(redactSensitiveInValue(JSON.parse(trimmed)));
+    } catch {
+      /* truncated / invalid JSON — fall through to the string scan */
+    }
+  }
+  return redactSensitiveInRawText(text);
+}
+
+/** Send-time pass: mask sensitive keys (built-in list + custom keys) across a
+ *  batch's bodies / frames / console args / custom props (mutates in place).
+ *  Always runs — the built-in list applies with no per-site config needed. */
+export function redactBatchSensitiveKeys(events: ObsrviqEvent[]): void {
+  for (const e of events) {
+    if (e.type === 'network' || e.type === 'ws') {
+      const anyE = e as unknown as {
+        requestBody?: { preview?: string };
+        responseBody?: { preview?: string };
+        body?: { preview?: string };
+      };
+      if (anyE.requestBody?.preview) anyE.requestBody.preview = redactSensitiveInText(anyE.requestBody.preview);
+      if (anyE.responseBody?.preview) anyE.responseBody.preview = redactSensitiveInText(anyE.responseBody.preview);
+      if (anyE.body?.preview) anyE.body.preview = redactSensitiveInText(anyE.body.preview);
+    } else if (e.type === 'custom') {
+      const c = e as unknown as { props?: Record<string, unknown> };
+      if (c.props) c.props = redactSensitiveInValue(c.props) as Record<string, unknown>;
+    } else if (e.type === 'console') {
+      const args = (e as unknown as { args?: Array<{ kind: string; json?: string }> }).args;
+      if (args) for (const a of args) if (a.kind === 'object' && a.json) a.json = redactSensitiveInText(a.json);
+    }
+  }
 }
 
 /** Recursively mask an object's string values and sensitive keys. */
@@ -36,7 +157,7 @@ export function maskValue(value: unknown, depth = 0): unknown {
   if (Array.isArray(value)) return value.map((v) => maskValue(v, depth + 1));
   const out: Record<string, unknown> = {};
   for (const [k, v] of Object.entries(value as Record<string, unknown>)) {
-    out[k] = isSensitiveKey(k) ? '«redacted»' : maskValue(v, depth + 1);
+    out[k] = isSensitiveKeyName(k) ? '«redacted»' : maskValue(v, depth + 1);
   }
   return out;
 }

package/src/network.ts

@@ -270,7 +270,8 @@ function wrapFetch(ctx: InstrumentCtx, tr: TimingResolver): Teardown {
   window.fetch = async function (input: RequestInfo | URL, init?: RequestInit) {
     const rawUrl = input instanceof Request ? input.url : String(input);
     // Never record our own telemetry round-trips to the ingest endpoint.
-    if (rawUrl.includes('/v1/batch')) return orig.call(window, input as RequestInfo, init);
+    if (rawUrl.includes('/v1/batch') || rawUrl.includes('/v1/config'))
+      return orig.call(window, input as RequestInfo, init);
     const startT = ctx.clock.now();
     const perfStart = perfNow();
     const method = (init?.method || (input instanceof Request ? input.method : 'GET')).toUpperCase();

package/src/transport.ts

@@ -1,5 +1,6 @@
 import { gzipSync } from 'fflate';
 import type { IngestBatch, IngestSessionMeta, ObsrviqEvent } from '@obsrviq/types';
+import { redactBatchSensitiveKeys, isMaskConfigPending } from './mask.js';
 
 export interface TransportOptions {
   endpoint: string;
@@ -55,10 +56,30 @@ export class Transport {
   /** Build, compress, and send the current buffer. Returns true on success. */
   async flush(reason: 'timer' | 'hidden' | 'pagehide' | 'manual' | 'backpressure'): Promise<boolean> {
     if (this.sending || this.buffer.length === 0) return true;
+
+    // Hold non-unload sends until the per-site mask config has loaded, so a
+    // remote-only custom key can't leak in the brief init fetch window. The fetch
+    // is bounded (index.ts marks config ready within ~2s even on failure/timeout),
+    // so this delays at most the first batch — it never blocks data indefinitely.
+    // Unload flushes (pagehide/hidden) can't wait, so they proceed best-effort.
+    const unloading = reason === 'pagehide' || reason === 'hidden';
+    if (!unloading && isMaskConfigPending()) {
+      if (!this.stopped) setTimeout(() => void this.flush('timer'), 200);
+      return true;
+    }
+
     this.sending = true;
     const events = this.buffer;
     this.buffer = [];
 
+    // Send-time safety net: mask sensitive keys across the batch — the built-in
+    // list (password/token/secret/…) PLUS any customer-defined keys — so a
+    // `{"password":"…"}` style body can't leave the page unmasked. Network/WS
+    // bodies get PII-pattern masking at capture but not key-based masking, so this
+    // is where key redaction lands for them; it also covers custom keys captured
+    // before the per-site config arrived.
+    redactBatchSensitiveKeys(events);
+
     let batch: IngestBatch | null = {
       siteKey: this.opts.siteKey,
       sessionId: this.opts.sessionId,

package/src/websocket.ts

@@ -0,0 +1,118 @@
+import type { WsEvent } from '@obsrviq/types';
+import type { InstrumentCtx, Teardown } from './context.js';
+import { maskString } from './mask.js';
+import { uuid } from './util.js';
+
+const FRAME_CAP = 4 * 1024; // 4 KB text-frame preview cap (mirrors network bodies)
+
+function redactUrl(raw: string): string {
+  try {
+    const u = new URL(raw, location.href);
+    if (u.search) u.search = '?…';
+    return u.toString();
+  } catch {
+    return raw.split('?')[0] ?? raw;
+  }
+}
+
+function byteLen(s: string): number {
+  try {
+    return new TextEncoder().encode(s).length;
+  } catch {
+    return s.length;
+  }
+}
+
+type FrameInfo = Pick<WsEvent, 'encoding' | 'size' | 'body'>;
+
+/** Derive encoding + byte size (+ opt-in masked preview) from a frame payload.
+ *  Never reads binary payload bytes — only their length. */
+function frameInfo(data: unknown, capture: boolean): FrameInfo {
+  if (typeof data === 'string') {
+    const info: FrameInfo = { encoding: 'text', size: byteLen(data) };
+    if (capture) {
+      const masked = maskString(data);
+      const truncated = masked.length > FRAME_CAP;
+      info.body = {
+        preview: truncated ? masked.slice(0, FRAME_CAP) : masked,
+        truncated,
+        redacted: masked !== data,
+      };
+    }
+    return info;
+  }
+  let size = 0;
+  if (typeof Blob !== 'undefined' && data instanceof Blob) size = data.size;
+  else if (data instanceof ArrayBuffer) size = data.byteLength;
+  else if (ArrayBuffer.isView(data)) size = (data as ArrayBufferView).byteLength;
+  return { encoding: 'binary', size };
+}
+
+/** Wrap the WebSocket constructor to capture connection lifecycle + frames.
+ *  Best-effort: never throws into the app; a capture failure never breaks send/recv. */
+export function instrumentWebSocket(ctx: InstrumentCtx): Teardown {
+  const Orig = typeof window !== 'undefined' ? window.WebSocket : undefined;
+  if (!Orig) return () => {};
+
+  const emit = (partial: Omit<WsEvent, 'id' | 'sessionId' | 'type' | 't' | 'ts'>) => {
+    try {
+      const ev: WsEvent = {
+        id: uuid(),
+        sessionId: ctx.sessionId,
+        type: 'ws',
+        t: ctx.clock.now(),
+        ts: ctx.clock.wall(),
+        ...partial,
+      };
+      ctx.emit(ev);
+    } catch {
+      /* swallow — capture must never affect the socket */
+    }
+  };
+
+  class WrappedWebSocket extends Orig {
+    constructor(url: string | URL, protocols?: string | string[]) {
+      super(url as string, protocols as string | string[] | undefined);
+      // The real socket is already live; capture setup must never throw into the app.
+      try {
+        const connId = uuid();
+        const redUrl = redactUrl(String(url));
+        const protos = protocols ? (Array.isArray(protocols) ? protocols : [protocols]) : undefined;
+        const capture = ctx.config.captureNetworkBodies;
+
+        this.addEventListener('open', () => emit({ connId, url: redUrl, kind: 'open', protocols: protos }));
+        this.addEventListener('close', (e) =>
+          emit({ connId, url: redUrl, kind: 'close', code: e.code, reason: e.reason || undefined, wasClean: e.wasClean }),
+        );
+        this.addEventListener('error', () => emit({ connId, url: redUrl, kind: 'error' }));
+        this.addEventListener('message', (e: MessageEvent) => {
+          const info = frameInfo(e.data, capture);
+          emit({ connId, url: redUrl, kind: 'message', dir: 'recv', ...info });
+          ctx.markActivity();
+        });
+
+        // Capture outgoing frames by shadowing send() on this instance. Call the
+        // real send FIRST — if it throws (e.g. send() while CONNECTING) the error
+        // propagates to the app and no phantom "send" frame is recorded.
+        const origSend = super.send.bind(this);
+        this.send = (data: string | ArrayBufferLike | Blob | ArrayBufferView) => {
+          origSend(data);
+          try {
+            const info = frameInfo(data, capture);
+            emit({ connId, url: redUrl, kind: 'message', dir: 'send', ...info });
+            ctx.markActivity();
+          } catch {
+            /* ignore capture errors */
+          }
+        };
+      } catch {
+        /* capture setup failed — the real socket still works normally. */
+      }
+    }
+  }
+
+  window.WebSocket = WrappedWebSocket as unknown as typeof WebSocket;
+  return () => {
+    window.WebSocket = Orig;
+  };
+}