@blamejs/core 0.15.64 → 0.15.66

package/CHANGELOG.md

@@ -8,6 +8,10 @@ upgrading across more than a few patches at a time.
 
 ## v0.15.x
 
+- v0.15.66 (2026-06-29) — **Recursive serializers and parsers refuse pathologically deep input with a typed error instead of overflowing the stack and crashing the process.** Several of the framework's recursive walkers over attacker-reachable input lacked an effective nesting cap, so a deeply nested value could exhaust the V8 call stack and throw an uncaught RangeError — crashing the process (a denial of service). b.canonicalJson is the most exposed: content-credentials verification canonicalises an untrusted manifest before it checks the signature, so a hostile credential could crash a verifier pre-authentication. b.jsonSchema.validate had a depth guard, but its limit was set so high the stack overflowed before the guard could fire, so validating a request body against a recursive schema crashed rather than returning a typed error. b.i18n.messageFormat parsed and rendered nested plural/select cases with no cap. Each walker now throws a typed framework error well before native overflow, and legitimate (even deeply nested) input is unaffected. The release also corrects DNSSEC signature-window comparison to the RFC 1982 serial-number arithmetic the spec requires, and makes the Redis client treat a malformed reply frame as a connection fault rather than letting it crash the host. **Fixed:** *DNSSEC compares RRSIG validity windows with RFC 1982 serial arithmetic* — b.network.dns.dnssec compared an RRSIG's inception and expiration against the current time by magnitude. RFC 4034 §3.1.5 requires the 32-bit timestamps to be interpreted with RFC 1982 serial-number arithmetic, which agrees with a plain comparison only while the values stay below 2^31. A signature whose window straddles the 2^31 (January 2038) or 2^32 (February 2106) boundary was mis-ordered — an in-window signature rejected, or a stale one accepted. The comparison now masks the clock into the same 32-bit serial space and tests the wrapped signed delta. **Security:** *Canonical JSON refuses excessively nested input before it can overflow the stack* — b.canonicalJson.stringify and stringifyJcs walk the value recursively. They detected circular references but had no nesting cap, so a deeply nested (acyclic) object or array overflowed the V8 stack with an uncaught RangeError. This is reachable before authentication: content-credentials verification canonicalises the untrusted manifest to hash it before verifying the signature, so a hostile credential could crash the verifier. Both serializers now throw a typed nesting-depth error at a depth far beyond any legitimate signed document and well short of native overflow. · *JSON Schema validation depth guard now fires before the stack overflows* — b.jsonSchema.validate caps subschema-validation nesting to defend against a recursive schema (for example items pointing back at the root with $ref) applied to a deeply nested instance — both attacker-controlled when validating a request body. The cap was set so high the V8 stack overflowed first, so the guard never ran and a crafted body crashed the validator with an uncaught RangeError instead of the typed reference-depth error. The limit is now well under native overflow; legitimate documents — deeply nested or wide — continue to validate. · *ICU MessageFormat refuses pathologically nested templates* — b.i18n.messageFormat.parse and format recurse once per nested plural/select case, with no depth cap. A template nested thousands of levels deep overflowed the stack. format() and parse() are public and b.i18n.t can render operator-supplied entries, so a hostile template is a denial-of-service vector; it now fails as a typed bad-template error. Real-world nesting (a handful of levels) is unaffected. · *Redis client treats a malformed reply frame as a connection fault, not a crash* — The RESP decoder recurses on nested arrays with no depth cap, and the data handler did not guard the parse, so a malformed or hostilely deep frame from a server threw out of the socket callback and could crash the host. The decoder now caps reply nesting and any parse fault tears the socket down and rejects in-flight commands for a reconnect — the same path as any other lost connection.
+
+- v0.15.65 (2026-06-29) — **The account-takeover kill-switch now actually locks the account: b.auth.lockout gains a lock() method and the kill-switch engages it through the operator's lockout instance.** b.auth.atoKillSwitch.trigger is the incident-response path for a compromised account: it destroys the user's sessions and then locks them out of new logins. The lockout step called b.auth.lockout.lock(), but no such method existed (and the kill-switch invoked it on the lockout module, which has no store), so the call threw and was swallowed — the kill-switch reported success while never locking the account, leaving an attacker who still held the credentials free to log straight back in. b.auth.lockout instances now expose lock(key, { durationMs?, untilMs?, reason? }) to force an account into lockout immediately, atomically, independent of the failure counter, and atoKillSwitch.trigger engages it through the lockout instance the operator passes as opts.lockout. When no lockout instance is supplied the step is skipped and the result's lockoutApplied is false (with an audit row), rather than silently claiming the account was locked. The release also closes a lock-clear race in lockout and refuses a self-disabling window. **Added:** *b.auth.lockout instances expose lock()* — lock(key, { durationMs?, untilMs?, reason? }) forces an account into lockout immediately — the operator action behind an ATO kill-switch or incident response — independent of the failure counter, defaulting to a long admin lock. A forced lock is released only by an explicit unlock(): recordSuccess() does not clear it, so a successful login by someone who still holds the compromised password cannot release a kill-switch lock. It uses the same atomic compare-and-set as the failure counter and throws if the lock cannot be committed (the caller must know it did not lock). **Security:** *Account-takeover kill-switch engages the lockout instead of silently doing nothing* — b.auth.atoKillSwitch.trigger invoked b.auth.lockout.lock(), which did not exist; the call threw and was caught by a best-effort guard, so the kill-switch destroyed sessions but never locked the account — an attacker still holding valid credentials could immediately re-authenticate. b.auth.lockout instances now provide lock() (an atomic, compare-and-set forced lockout), and the kill-switch calls it on the lockout instance supplied as opts.lockout. Without a lockout instance the lockout step cannot run; the result's lockoutApplied is false and an audit row records that the account was not locked, so an operator is not misled into believing a compromised account was secured. opts.lockout is now the lockout instance (or false to skip), not a boolean toggle. · *lockout clears the failure counter atomically and refuses a zero window* — b.auth.lockout.recordSuccess and unlock cleared state with a read-then-delete, which could erase a lock a concurrent recordFailure had just engaged; they now clear under the same compare-and-set the failure path uses. create() also refuses windowMs: 0, which previously disabled lockout entirely (every failure decayed immediately and the zero-TTL state never persisted).
+
 - v0.15.64 (2026-06-29) — **Token and HTTP-message-signature verifiers now reject a non-finite clock-skew or tolerance value instead of letting it disable the expiry, not-before, and future-dating checks.** Several verifiers read an operator-supplied clock-skew or tolerance value with a bare type check that accepted Infinity and NaN. Because the resulting comparison (for example exp + skew < now, or age > tolerance) is always false when the skew or tolerance is Infinity, a misconfigured non-finite value silently disabled the time-window check — so an expired token, a not-yet-valid token, a future-dated signature, or a replayed (too-old) signature would verify. The affected paths are the shared external-JWT verifier b.auth.jwt.verifyExternal (which the JAR / signed-request-object path also uses), the SD-JWT-VC credential verifier, the OAuth ID-token and client-attestation verifiers, and the HTTP message signature verifier's freshness and clock-skew windows. Each now requires a present skew or tolerance to be a non-negative finite number: the throw-based JWT, SD-JWT-VC, and OAuth verifiers reject a malformed value, and the verdict-returning HTTP-message-signature verifier falls back to its safe default rather than disabling the window. A build check now fails if any verifier reads one of these options without a finiteness guard. **Added:** *verifyExternal and jar.parse accept an `now` clock override* — b.auth.jwt.verifyExternal and b.auth.jar.parse now accept an optional `now` (a finite epoch-milliseconds value) to evaluate a token's exp / nbf / iat as of a specific instant instead of the wall clock — consistent with b.auth.sdJwtVc.verify. A negative clockSkewMs is no longer accepted as a way to shift the evaluation time; use `now`. **Security:** *Non-finite clock-skew / tolerance no longer disables a verifier's time-window check* — b.auth.jwt.verifyExternal, b.auth.sdJwtVc.verify, b.auth.oauth (ID-token and client-attestation verification), and the HTTP message signature verifier read their clock-skew, max-clock-skew, max-PoP-age, and tolerance options with a bare `typeof === "number"` check that let Infinity and NaN through. With such a value the expiry / not-before / future-dating comparison is always false, so the check was silently skipped and an expired or not-yet-valid token — or a future-dated or replayed signature — would verify. The JWT, SD-JWT-VC, and OAuth verifiers now reject a non-finite or negative skew (the option is operator configuration, not attacker-controlled), and the HTTP message signature verifier falls back to its default tolerance and skew. The CWT, DPoP, SAML, and OpenID Federation verifiers already enforced this and are unchanged. **Detectors:** *Verifier clock-skew / tolerance options must be finite-guarded* — A build check fails if a verifier reads a clock-skew or tolerance option with a bare `typeof === "number"` without a finiteness guard (a non-negative-finite check, an inline isFinite fallback, or a create-time non-negative-finite schema), preventing a future verifier from reintroducing the disable-the-window class.
 
 - v0.15.63 (2026-06-29) — **OID4VCI now enforces single-use of a pre-authorized code and of a single-use access token under concurrency, so two simultaneous requests can no longer mint two credentials from one.** On the OID4VCI credential issuer, two single-use values were consumed by a delete whose result was ignored, so concurrent requests could each act on the same value. exchangePreAuthorizedCode read the pre-authorized code's entry, validated the transaction code, then deleted the code and minted an access token without checking that its own call had removed the entry — two simultaneous /token requests with the same code each saw the entry, each deleted it, and each minted a distinct access token, issuing two access tokens (and ultimately two credentials) from a code OID4VCI requires to be single-use. issueCredential had the same shape: with single-use access tokens (the default), it minted the credential first and deleted the access token afterward as best-effort cleanup, so two concurrent requests bearing the same token both read it and the same not-yet-rotated c_nonce, both proofs verified, and both minted a credential. Both paths now claim the value with an atomic delete and proceed only when that delete succeeded; the losing request is refused. The transaction-code and proof checks still run first, so a bad transaction code or proof does not consume the value (a wallet can retry). **Security:** *OID4VCI pre-authorized code and access token are single-use under concurrency* — exchangePreAuthorizedCode and issueCredential consumed their single-use value (the pre-authorized code, and the single-use access token) with a delete whose return was discarded, and in issueCredential's case after the credential was already minted. Two concurrent requests carrying the same value therefore each succeeded — minting two access tokens from one pre-authorized code, or two credentials from one single-use access token — defeating the single-use guarantee OID4VCI §3.5 requires (an authorization intended for one credential could yield two). Both paths now delete the value atomically and issue only if that delete removed it, refusing the request that lost the race; the transaction-code and proof verifications run before the claim, so an invalid attempt does not burn the value. If the operator's credential issuer throws after the access token has been claimed (a transient signer outage), the token is restored so the wallet can retry rather than being permanently consumed without a credential.

package/lib/auth/ato-kill-switch.js

@@ -10,17 +10,22 @@
  * cleanup path once the trigger fires:
  *
  *   1. destroy every session for the user across the cluster
- *   2. lock the user out of new logins (b.auth.lockout)
+ *   2. lock the user out of new logins via the supplied lockout instance
  *   3. emit an audit row with reason / actor for downstream forensics
  *
  *   await b.auth.atoKillSwitch.trigger({
  *     userId:    "u_42",
  *     reason:    "fraud-signal: 14 failed MFA from new geo",
  *     actor:     { id: req.user && req.user.id, role: req.user && req.user.role },
- *     lockout:   true,                  // default true
+ *     lockout:   appLockout,            // a b.auth.lockout instance (or false to skip)
  *     accessLock: "locked",             // optional — flip the global access-lock mode
  *   });
  *
+ * The lockout step needs the operator's configured b.auth.lockout instance
+ * (there is no module-level lockout store); pass it as `opts.lockout`. When it
+ * is omitted, the lockout step is skipped and the returned `lockoutApplied` is
+ * false (with an audit row) — the account is NOT locked.
+ *
  * Returns `{ sessionsDestroyed, lockoutApplied, accessLockMode }`.
  */
 
@@ -29,7 +34,6 @@ var validateOpts = require("../validate-opts");
 var { defineClass } = require("../framework-error");
 
 var session = lazyRequire(function () { return require("../session"); });
-var lockout = lazyRequire(function () { return require("./lockout"); });
 var accessLock = lazyRequire(function () { return require("./access-lock"); });
 var audit = lazyRequire(function () { return require("../audit"); });
 
@@ -43,8 +47,15 @@ async function trigger(opts) {
 
   validateOpts.requireNonEmptyString(opts.userId, "userId", AtoKillSwitchError, "auth-ato-kill-switch/missing-user-id");
   validateOpts.requireNonEmptyString(opts.reason, "reason", AtoKillSwitchError, "auth-ato-kill-switch/missing-reason");
-  var doLockout       = opts.lockout !== false;
-  var accessLockMode  = typeof opts.accessLock === "string" ? opts.accessLock : null;
+  // opts.lockout is the operator's configured b.auth.lockout INSTANCE to lock
+  // the account in (or false to skip the lockout step). The kill-switch needs
+  // that instance's store to engage the lock — there is no module-level lockout
+  // singleton — so without an instance the lockout step cannot run and is
+  // reported (rather than silently claiming success while never locking).
+  var skipLockout    = opts.lockout === false;
+  var lockoutInst    = (opts.lockout && typeof opts.lockout === "object" &&
+                        typeof opts.lockout.lock === "function") ? opts.lockout : null;
+  var accessLockMode = typeof opts.accessLock === "string" ? opts.accessLock : null;
 
   var sessionsDestroyed = 0;
   try {
@@ -63,13 +74,37 @@ async function trigger(opts) {
   }
 
   var lockoutApplied = false;
-  if (doLockout) {
-    try {
-      await lockout().lock(opts.userId, {
-        reason: "ato-kill-switch:" + opts.reason,
+  if (!skipLockout) {
+    if (lockoutInst) {
+      try {
+        await lockoutInst.lock(opts.userId, {
+          reason: "ato-kill-switch:" + opts.reason,
+        });
+        lockoutApplied = true;
+      } catch (e) {
+        // The lockout step failed (e.g. a cache outage) AFTER sessions were
+        // destroyed. Don't fail the whole kill-switch — but surface it so the
+        // operator knows the account was NOT locked, rather than swallowing.
+        audit().safeEmit({
+          action: "auth.ato_kill_switch.partial",
+          outcome: "failure",
+          metadata: { userId: opts.userId, step: "lockout", reason: e && e.message },
+        });
+      }
+    } else {
+      // No lockout instance supplied: the kill-switch cannot lock the account.
+      // Report it (the result's lockoutApplied stays false) so the operator
+      // doesn't believe the user was locked out of new logins.
+      audit().safeEmit({
+        action: "auth.ato_kill_switch.partial",
+        outcome: "failure",
+        metadata: {
+          userId: opts.userId,
+          step:   "lockout",
+          reason: "no lockout instance supplied (pass opts.lockout = b.auth.lockout.create({ cache, namespace }))",
+        },
       });
-      lockoutApplied = true;
-    } catch (_e) { /* lockout is best-effort; sessions already destroyed */ }
+    }
   }
 
   var modeApplied = null;

package/lib/auth/lockout.js

@@ -90,6 +90,11 @@ var observability = lazyRequire(function () { return require("../observability")
 
 var _err = LockoutError.factory;
 
+// Default duration for an operator-forced lock() (ATO kill-switch / incident
+// response) when neither untilMs nor durationMs is supplied — long enough to
+// require an explicit admin unlock() during an active incident.
+var DEFAULT_ADMIN_LOCK_MS = C.TIME.hours(24);
+
 var DEFAULTS = Object.freeze({
   maxAttempts:     5,
   windowMs:        C.TIME.minutes(15),
@@ -173,7 +178,11 @@ function create(opts) {
   _requirePositiveInt("maxAttempts", maxAttempts);
 
   var windowMs = opts.windowMs !== undefined ? opts.windowMs : DEFAULTS.windowMs;
-  _requireNonNegFinite("windowMs", windowMs);
+  // windowMs must be POSITIVE: a 0 window makes every failure "decay" on the
+  // next request (the `now - lastFailureAt > windowMs` reset fires for any
+  // elapsed time) so the counter never reaches maxAttempts, AND the cache TTL
+  // of 0 makes the state non-persistent — together silently disabling lockout.
+  _requirePositiveInt("windowMs", windowMs);
 
   var lockoutDurations = opts.lockoutDurations !== undefined
                             ? opts.lockoutDurations : DEFAULTS.lockoutDurations;
@@ -242,6 +251,49 @@ function create(opts) {
     }
   }
 
+  // Atomically clear the lockout state under a compare-and-set, so a clear
+  // (recordSuccess / unlock) cannot race a concurrent recordFailure that just
+  // engaged a lock — a read-then-del would erase that fresh lock. onState is
+  // invoked with the pre-clear state inside the CAS mutator (it may re-run on a
+  // retry; the last invocation reflects the committed state) to capture audit
+  // detail. On a backend that can't do an atomic update at runtime it falls
+  // back to read-then-del (a lost clear leaves a lock in place — fail-safe).
+  // preserveIf(state) → true to KEEP the state (abort the clear) rather than
+  // delete it. recordSuccess passes a predicate that preserves an active forced
+  // (admin / ATO kill-switch) lock, so a successful login by someone who still
+  // holds the compromised password cannot clear it — only an explicit unlock()
+  // releases a forced lock.
+  async function _atomicClear(key, onState, preserveIf) {
+    try {
+      await cache.update(_scopedKey(key), function (state) {
+        state = state || null;
+        onState(state);
+        if (!state) return { abort: true };
+        if (preserveIf && preserveIf(state)) return { abort: true };
+        return { delete: true };
+      }, { ttlMs: windowMs });
+    } catch (e) {
+      if (e && e.code === "UNSUPPORTED") {
+        var st = await _readState(key);
+        onState(st);
+        if (st && !(preserveIf && preserveIf(st))) await _deleteState(key);
+        return;
+      }
+      _signalCacheError("update");
+      // Best-effort fallback so a transient cache error doesn't leave the
+      // counter un-cleared on the happy path.
+      var st2 = await _readState(key);
+      onState(st2);
+      if (st2 && !(preserveIf && preserveIf(st2))) await _deleteState(key);
+    }
+  }
+
+  // An active forced/admin lock (set by lock()) must survive recordSuccess.
+  function _isActiveForcedLock(state) {
+    return !!(state && state.forced === true &&
+              typeof state.lockedUntil === "number" && state.lockedUntil > clock());
+  }
+
   function _verdictFromState(state, now) {
     if (!state) return { locked: false, attempts: 0 };
     if (state.lockedUntil && state.lockedUntil > now) {
@@ -380,13 +432,19 @@ function create(opts) {
   async function recordSuccess(key, callOpts) {
     _requireKey(key);
     callOpts = callOpts || {};
-    var state = await _readState(key);
-    var hadCounter = !!(state && (state.attempts > 0 || state.lockedUntil));
-    if (state) await _deleteState(key);
+    var hadCounter = false;
+    var clearedAttempts = 0;
+    // Preserve an active forced (admin / ATO) lock: a successful credential
+    // verification must not release a kill-switch lock — the password may still
+    // be the compromised one. A forced lock yields only to unlock().
+    await _atomicClear(key, function (state) {
+      hadCounter = !!(state && (state.attempts > 0 || state.lockedUntil));
+      clearedAttempts = (state && state.attempts) || 0;
+    }, _isActiveForcedLock);
     _emitObs("auth.lockout.success", { namespace: namespace });
     if (auditSuccess) {
       _emitAudit("auth.lockout.success", key, "success",
-        { attemptsCleared: (state && state.attempts) || 0,
+        { attemptsCleared: clearedAttempts,
           hadCounter:      hadCounter },
         callOpts.req);
     }
@@ -401,26 +459,93 @@ function create(opts) {
   async function unlock(key, callOpts) {
     _requireKey(key);
     callOpts = callOpts || {};
-    var state = await _readState(key);
     var now = clock();
-    var hadLock = !!(state && (
-      (state.lockedUntil && state.lockedUntil > now) ||
-      (state.attempts || 0) > 0
-    ));
-    if (state) await _deleteState(key);
+    var hadLock = false;
+    var prior = { attempts: 0, lockedUntil: null, lockNumber: 0 };
+    await _atomicClear(key, function (state) {
+      hadLock = !!(state && (
+        (state.lockedUntil && state.lockedUntil > now) ||
+        (state.attempts || 0) > 0
+      ));
+      prior = {
+        attempts:    (state && state.attempts) || 0,
+        lockedUntil: (state && state.lockedUntil) || null,
+        lockNumber:  (state && state.lockNumber) || 0,
+      };
+    });
     _emitObs("auth.lockout.unlock", { namespace: namespace });
     if (auditUnlock) {
       _emitAudit("auth.lockout.unlock", key, "success",
         { hadLock:           hadLock,
-          priorAttempts:    (state && state.attempts) || 0,
-          priorLockedUntil: (state && state.lockedUntil) || null,
-          priorLockNumber:  (state && state.lockNumber) || 0,
+          priorAttempts:    prior.attempts,
+          priorLockedUntil: prior.lockedUntil,
+          priorLockNumber:  prior.lockNumber,
           reason:           callOpts.reason || null },
         callOpts.req);
     }
     return hadLock;
   }
 
+  // Force an account into lockout immediately — the operator action behind an
+  // ATO kill-switch / incident response, independent of the failure counter.
+  // Sets lockedUntil to `untilMs` (absolute) or now+`durationMs`, defaulting to
+  // a long admin lock; bumps lockNumber so a subsequent failure ladders from
+  // here. Uses the same compare-and-set as recordFailure (atomic, retried). An
+  // admin lock that cannot be committed THROWS (the caller — e.g. the kill-
+  // switch — must know it did not lock), rather than the hot-path fail-open.
+  async function lock(key, callOpts) {
+    _requireKey(key);
+    callOpts = callOpts || {};
+    var now = clock();
+    var lockedUntil;
+    if (typeof callOpts.untilMs === "number" && isFinite(callOpts.untilMs)) {
+      lockedUntil = callOpts.untilMs;
+    } else if (typeof callOpts.durationMs === "number" && isFinite(callOpts.durationMs) && callOpts.durationMs > 0) {
+      lockedUntil = now + callOpts.durationMs;
+    } else {
+      lockedUntil = now + DEFAULT_ADMIN_LOCK_MS;
+    }
+    if (lockedUntil <= now) {
+      throw _err("BAD_OPT", "lock: resolved lockedUntil is not in the future " +
+        "(untilMs/durationMs) — use unlock() to clear a lock");
+    }
+    var ttl = lockedUntil - now + windowMs;
+    var lockNumber = 0;
+    try {
+      await cache.update(_scopedKey(key), function (state) {
+        lockNumber = ((state && state.lockNumber) || 0) + 1;
+        return {
+          value: {
+            attempts:       0,
+            lockNumber:     lockNumber,
+            firstFailureAt: (state && state.firstFailureAt) || now,
+            lastFailureAt:  now,
+            lockedUntil:    lockedUntil,
+            // Marks this as an operator-forced lock so recordSuccess won't clear
+            // it — it is released only by an explicit unlock().
+            forced:         true,
+          },
+          ttlMs: ttl,
+        };
+      }, { ttlMs: ttl });
+    } catch (e) {
+      if (e && e.code === "UNSUPPORTED") {
+        throw _err("CACHE_NO_ATOMIC_UPDATE",
+          "auth.lockout: the cache backend does not support atomic update() — " +
+          "lock() cannot enforce the lockout across nodes on a get/set-only backend.");
+      }
+      throw e;
+    }
+    _emitObs("auth.lockout.engaged", { namespace: namespace, lockNumber: String(lockNumber) });
+    if (auditEngaged) {
+      _emitAudit("auth.lockout.engaged", key, "denied",
+        { lockNumber: lockNumber, lockedUntil: lockedUntil, durationMs: lockedUntil - now,
+          forced: true, reason: callOpts.reason || null },
+        callOpts.req);
+    }
+    return { locked: true, lockedUntil: lockedUntil, lockNumber: lockNumber };
+  }
+
   async function attempts(key) {
     _requireKey(key);
     var state = await _readState(key);
@@ -437,6 +562,7 @@ function create(opts) {
     recordFailure: recordFailure,
     recordSuccess: recordSuccess,
     check:         check,
+    lock:          lock,
     unlock:        unlock,
     attempts:      attempts,
     close:         close,

package/lib/canonical-json.js

@@ -42,7 +42,17 @@
 // RFC 8785 §3.2.3 sort. Primitives, strings, and numbers use
 // JSON.stringify, whose escaping (§3.2.2.2) and ECMAScript number format
 // (§3.2.2.3) are exactly what JCS references.
-function _emit(value, seen, bufferAs) {
+
+// Maximum nesting depth. The walk is recursive, so a deeply-nested input
+// (e.g. an attacker-supplied manifest handed to content-credentials.verify
+// BEFORE signature verification, or untrusted MCP tool-call args) would
+// otherwise overflow the V8 stack with an unhandled RangeError — a crash /
+// pre-auth DoS. Throw a typed framework error well before native overflow.
+// The cycle WeakSet below catches references that loop; this catches a tree
+// that is merely very deep. 512 is far beyond any legitimate signed document.
+var MAX_DEPTH = 512;
+
+function _emit(value, seen, bufferAs, depth) {
   if (value === null || typeof value === "undefined") return "null";
   var t = typeof value;
   if (t === "number" || t === "string" || t === "boolean") return JSON.stringify(value);
@@ -90,6 +100,10 @@ function _emit(value, seen, bufferAs) {
   if (seen.has(value)) {
     throw new Error("canonical-json: circular reference detected");
   }
+  depth = depth || 0;
+  if (depth > MAX_DEPTH) {
+    throw new Error("canonical-json: maximum nesting depth exceeded (" + MAX_DEPTH + ")");
+  }
   seen.add(value);
   if (Array.isArray(value)) {
     // Index loop, not .map(): map() skips holes in a sparse array,
@@ -97,7 +111,7 @@ function _emit(value, seen, bufferAs) {
     // reads as undefined → _emit returns "null" (matching JSON.stringify).
     var items = [];
     for (var ai = 0; ai < value.length; ai += 1) {
-      items.push(_emit(value[ai], seen, bufferAs));
+      items.push(_emit(value[ai], seen, bufferAs, depth + 1));
     }
     return "[" + items.join(",") + "]";
   }
@@ -107,7 +121,7 @@ function _emit(value, seen, bufferAs) {
   keys.sort();
   var parts = [];
   for (var i = 0; i < keys.length; i++) {
-    parts.push(JSON.stringify(keys[i]) + ":" + _emit(value[keys[i]], seen, bufferAs));
+    parts.push(JSON.stringify(keys[i]) + ":" + _emit(value[keys[i]], seen, bufferAs, depth + 1));
   }
   return "{" + parts.join(",") + "}";
 }

package/lib/i18n-messageformat.js

@@ -63,6 +63,18 @@ function _err(code, message) {
   return new I18nMessageFormatError(code, message, true);
 }
 
+// Maximum case-nesting depth. The parser and the renderer both recurse
+// once per nested plural/select case body; a template like
+// `{a,select,x{{b,select,x{{c,select,...}}}}}` nested thousands deep would
+// otherwise overflow the V8 stack with an uncaught RangeError. Templates
+// usually come from operator translation files, but format()/parse() are
+// public and `b.i18n.t(key, vars, { messageFormat: true })` renders entries
+// that may be operator-supplied per-tenant, so a hostile template must fail
+// as a typed BAD_TEMPLATE rather than crashing the process. Real
+// translations nest 2-3 deep; 100 is far beyond any legitimate message yet
+// well under native overflow.
+var MAX_NESTING_DEPTH = 100;
+
 // ---- Parser ----
 //
 // AST node shapes:
@@ -89,6 +101,12 @@ function parse(template) {
 }
 
 function _parseSequence(state, topLevel) {
+  state.depth = (state.depth || 0) + 1;
+  if (state.depth > MAX_NESTING_DEPTH) {
+    throw _err("BAD_TEMPLATE",
+      "messageFormat.parse: case nesting too deep (max " +
+      MAX_NESTING_DEPTH + ")");
+  }
   var nodes = [];
   var lit = "";
   while (state.pos < state.src.length) {
@@ -146,6 +164,7 @@ function _parseSequence(state, topLevel) {
     state.pos += 1;
   }
   if (lit.length > 0) nodes.push({ type: "literal", value: lit });
+  state.depth -= 1;
   return nodes;
 }
 
@@ -322,18 +341,26 @@ function _pluralRules(locale, type) {
 
 function format(template, vars, locale) {
   var nodes = parse(template);
-  return _renderSequence(nodes, vars || {}, locale || "en", null);
+  return _renderSequence(nodes, vars || {}, locale || "en", null, 0);
 }
 
-function _renderSequence(nodes, vars, locale, hashContext) {
+function _renderSequence(nodes, vars, locale, hashContext, depth) {
+  depth = depth || 0;
+  // parse() already bounds AST nesting to MAX_NESTING_DEPTH, so this guard
+  // is defence-in-depth for any future caller that hands render a hand-built
+  // tree — it must still fail typed rather than overflow the stack.
+  if (depth > MAX_NESTING_DEPTH) {
+    throw _err("BAD_TEMPLATE",
+      "messageFormat: render nesting too deep (max " + MAX_NESTING_DEPTH + ")");
+  }
   var out = "";
   for (var i = 0; i < nodes.length; i++) {
-    out += _renderNode(nodes[i], vars, locale, hashContext);
+    out += _renderNode(nodes[i], vars, locale, hashContext, depth);
   }
   return out;
 }
 
-function _renderNode(node, vars, locale, hashContext) {
+function _renderNode(node, vars, locale, hashContext, depth) {
   if (node.type === "literal") return node.value;
   if (node.type === "hash") {
     return hashContext != null ? String(hashContext) : "#";
@@ -358,13 +385,13 @@ function _renderNode(node, vars, locale, hashContext) {
       var category = pr.select(adjusted);
       caseBody = node.cases[category] || node.cases.other;
     }
-    return _renderSequence(caseBody, vars, locale, adjusted);
+    return _renderSequence(caseBody, vars, locale, adjusted, depth + 1);
   }
   if (node.type === "select") {
     var sv = vars[node.name];
     var key = (sv === undefined || sv === null) ? "other" : String(sv);
     var body = node.cases[key] || node.cases.other;
-    return _renderSequence(body, vars, locale, hashContext);
+    return _renderSequence(body, vars, locale, hashContext, depth + 1);
   }
   return "";
 }

package/lib/json-schema.js

@@ -54,7 +54,17 @@ var { defineClass } = require("./framework-error");
 var JsonSchemaError = defineClass("JsonSchemaError", { alwaysPermanent: true });
 
 var DIALECT_2020_12 = "https://json-schema.org/draft/2020-12/schema";
-var MAX_REF_DEPTH = 10000;                                  // recursion-depth cap (count, not a byte size)
+// Subschema-validation nesting cap. ctx.depth is incremented on entry to
+// _validate and decremented in its finally, so it tracks the live nesting
+// depth (sibling keywords/properties do not accumulate) — the deepest the
+// instance×schema walk has recursed. validate(schema, requestBody) reaches
+// it with attacker input on BOTH sides: a recursive schema (items:{$ref:"#"})
+// against a deeply-nested array/object, or a deeply-nested allOf chain.
+// Left unbounded the walk overflows the V8 stack (~1000 levels) with an
+// uncaught RangeError before this guard fired — a pre-validation DoS. 256 is
+// far beyond any legitimate document yet well under native overflow, so the
+// typed json-schema/ref-loop error is what an operator sees.
+var MAX_REF_DEPTH = 256;                                    // recursion-depth cap (count, not a byte size)
 var DEFAULT_MAX_ERRORS = 100;                               // error-collection cap
 
 function _typeOf(v) {

package/lib/network-dnssec.js

@@ -295,8 +295,19 @@ function verifyRrset(opts) {
   validateOpts.optionalDate(opts.at, "dnssec.verifyRrset: opts.at", DnssecError, "dnssec/bad-at");
   var atMs = (opts.at !== undefined && opts.at !== null) ? opts.at.getTime() : Date.now();
   var nowSec = Math.floor(atMs / 1000);
-  if (nowSec < (rrsig.inception >>> 0)) throw new DnssecError("dnssec/not-yet-valid", "dnssec.verifyRrset: RRSIG inception is in the future");
-  if (nowSec > (rrsig.expiration >>> 0)) throw new DnssecError("dnssec/expired", "dnssec.verifyRrset: RRSIG has expired");
+  // RFC 4034 §3.1.5: the RRSIG inception/expiration fields are 32-bit and
+  // MUST be compared with RFC 1982 serial-number arithmetic, not magnitude.
+  // A plain `<` / `>` agrees with serial arithmetic only while both operands
+  // stay under 2^31; it mis-orders any window that straddles the 2^31 (Jan
+  // 2038) or 2^32 (Feb 2106) boundary — accepting an expired signature or
+  // rejecting a valid one. Mask the clock into the same 32-bit serial space
+  // and compare the wrapped signed delta: a negative (now - inception) means
+  // inception is in the future; a negative (expiration - now) means expired.
+  var nowSer = nowSec >>> 0;
+  var incepSer = rrsig.inception >>> 0;
+  var expirSer = rrsig.expiration >>> 0;
+  if (((nowSer - incepSer) | 0) < 0) throw new DnssecError("dnssec/not-yet-valid", "dnssec.verifyRrset: RRSIG inception is in the future");
+  if (((expirSer - nowSer) | 0) < 0) throw new DnssecError("dnssec/expired", "dnssec.verifyRrset: RRSIG has expired");
 
   var klass = typeof opts.class === "number" ? opts.class : 1;
   var ownerWire = _canonicalName(opts.name);

package/lib/redis-client.js

@@ -78,7 +78,20 @@ function _encodeCommand(args) {
 //   { type: "int",    value, consumed }   — integer (:42)
 //   { type: "bulk",   value, consumed }   — bulk string buffer (or null)
 //   { type: "array",  value, consumed }   — array of decoded items
-function _parseFrame(buf, offset) {
+//
+// RESP arrays nest (an array whose elements are arrays), so the decoder
+// recurses. A hostile or compromised server can stream an arbitrarily deep
+// nest of `*1\r\n` headers to overflow the V8 stack with an uncaught
+// RangeError out of the socket 'data' handler — a crash. Cap the nesting
+// well above any real reply (cluster-slots / XRANGE replies nest a handful
+// deep) and throw a typed PROTOCOL error, which _onData turns into a socket
+// teardown + reconnect instead of a process crash.
+var MAX_RESP_DEPTH = 64;
+function _parseFrame(buf, offset, depth) {
+  depth = depth || 0;
+  if (depth > MAX_RESP_DEPTH) {
+    throw _err("PROTOCOL", "reply nesting exceeds " + MAX_RESP_DEPTH + " levels");
+  }
   if (offset >= buf.length) return { type: "incomplete" };
   var marker = buf[offset];
   // Find next CRLF after the marker
@@ -121,7 +134,7 @@ function _parseFrame(buf, offset) {
     var items = [];
     var cursor = crlf + 2;
     for (var i = 0; i < arrLen; i++) {
-      var sub = _parseFrame(buf, cursor);
+      var sub = _parseFrame(buf, cursor, depth + 1);
       if (sub.type === "incomplete") return { type: "incomplete" };
       items.push(sub);
       cursor += sub.consumed;
@@ -299,9 +312,19 @@ function create(opts) {
   function _onData(chunk) {
     rxBuffer = rxBuffer.length === 0 ? chunk : Buffer.concat([rxBuffer, chunk]);
     while (rxBuffer.length > 0) {
-      var frame = _parseFrame(rxBuffer, 0);
-      if (frame.type === "incomplete") return;
-      var value = _frameToValue(frame);
+      var frame, value;
+      try {
+        frame = _parseFrame(rxBuffer, 0);
+        if (frame.type === "incomplete") return;
+        value = _frameToValue(frame);
+      } catch (parseErr) {
+        // A malformed or hostilely-nested RESP frame must not throw out of
+        // the socket 'data' handler and crash the host. Treat it as a fatal
+        // connection fault: reject in-flight commands and tear the socket
+        // down for a reconnect, the same as any other lost-socket path.
+        _teardownSocket(parseErr);
+        return;
+      }
       rxBuffer = rxBuffer.slice(frame.consumed);
 
       // Pub/sub push detection — server-initiated arrays beginning with

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@blamejs/core",
-  "version": "0.15.64",
+  "version": "0.15.66",
   "description": "The Node framework that owns its stack.",
   "license": "Apache-2.0",
   "author": "blamejs contributors",

package/sbom.cdx.json

@@ -2,10 +2,10 @@
   "$schema": "http://cyclonedx.org/schema/bom-1.5.schema.json",
   "bomFormat": "CycloneDX",
   "specVersion": "1.5",
-  "serialNumber": "urn:uuid:f9782178-0584-49aa-8fd6-3029b774391a",
+  "serialNumber": "urn:uuid:38fe8005-1751-4103-b424-551b80f8f722",
   "version": 1,
   "metadata": {
-    "timestamp": "2026-06-29T22:06:28.985Z",
+    "timestamp": "2026-06-30T00:16:16.691Z",
     "lifecycles": [
       {
         "phase": "build"
@@ -19,14 +19,14 @@
       }
     ],
     "component": {
-      "bom-ref": "@blamejs/core@0.15.64",
+      "bom-ref": "@blamejs/core@0.15.66",
       "type": "application",
       "name": "blamejs",
-      "version": "0.15.64",
+      "version": "0.15.66",
       "scope": "required",
       "author": "blamejs contributors",
       "description": "The Node framework that owns its stack.",
-      "purl": "pkg:npm/%40blamejs/core@0.15.64",
+      "purl": "pkg:npm/%40blamejs/core@0.15.66",
       "properties": [],
       "externalReferences": [
         {
@@ -54,7 +54,7 @@
   "components": [],
   "dependencies": [
     {
-      "ref": "@blamejs/core@0.15.64",
+      "ref": "@blamejs/core@0.15.66",
       "dependsOn": []
     }
   ]