Offline evidence checker for CVE-2026-56846 — Node.js: HTTP/2 retained header memory exhaustion. It compares operator-reported Node.js runtime versions with the fixed releases named in the vendor's security release and writes a conservative, version-only triage report. Python 3.10+, standard library only; the executable can be copied alone.
Version evidence, not an exposure test. Results describe the versions you report. They do not establish that a vulnerable code path is reachable, that a deployed runtime has been patched, or that anything was exploited.
- Scope
- Requirements
- Quick start
- Synthetic report preview
- Usage
- Fixed versions and classification rules
- Interpreting results
- Report formats and provenance
- Limits, exit codes and interruption
- Sources
- Project documents
- Contributing and security
- License
The checker:
- compares versions you supply (
--version, a JSON inventory or a CSV list) with the fixed releases for Node.js 22.x and 24.x; - writes CSV (the default), JSON or script-free HTML reports that all carry the same evidence.
The checker never:
- scans the current machine, contacts websites or any network service, executes Node.js or project code, or sends exploit payloads;
- tests exploitability: every result carries
exploitability: not_tested; - attests that a deployed runtime is patched. A reported fixed version is a claim by the operator, not proof of what is running.
Hidden versions, custom builds, conflicting evidence and uncovered release lines remain unknown. A public URL is not an inventory.
- Python 3.10 or later and nothing else: no third-party packages and no network access at runtime. CI runs the tests on Python 3.10 and 3.14 on Ubuntu and Windows.
- The exact Node.js version obtained from the deployed process or container, and its build source when the build is not an upstream release.
- Authorization to assess the systems your inventories describe. The tool has no network mode, and running it authorizes nothing.
# Assess one reported version and print JSON
python node_cve_2026_56846_check.py --version 22.23.2 --format json
# Assess the synthetic example inventory and save an HTML report
python node_cve_2026_56846_check.py --inventory examples/inventory.example.json --format html --output report.html
# Assess a CSV list (CSV is the default format) and exit 1 on a finding
python node_cve_2026_56846_check.py --csv-list inventory.csv --fail-on-finding
# Show all options, then run the local-only test suite
python node_cve_2026_56846_check.py --help
python -m unittest -vThere is no automatic scan of the current machine: you supply every version. --version is repeatable; each value describes a separate runtime.
Open the synthetic example report. Example data only; this is not a live assessment.
Provide at least one input: --version, --inventory or --csv-list. Inputs can be combined in one run, and every input produces its own result row.
| Option | Purpose |
|---|---|
--version VERSION (alias --node-version) |
Assess an explicitly supplied Node.js version offline. A leading v is accepted. Repeatable; each value is a separate runtime. |
--build-source SOURCE |
Provenance of the --version builds, at most 512 characters. unknown (the default), upstream, official and official_upstream are compared as upstream releases (matched case-insensitively); any other value, such as a vendor name, yields unknown with vendor_build_unverified. |
--http2-service {true,false} |
Operator observation applied to every --version input. |
--untrusted-http2-clients {true,false} |
Operator observation applied to every --version input. |
--inventory FILE |
JSON object or array of entries. Repeatable. |
--csv-list FILE |
CSV list with a node_version column. Repeatable. |
--format {csv,json,html} |
Report format. Default: csv. |
--output FILE |
Save the report to a new file in an existing directory. Default: stdout. |
--force |
Replace an existing output file. Inputs can never be overwritten, even with --force. |
--brand NAME |
Organization name shown in the HTML report, at most 512 characters. Default: CompleteTech LLC. |
--fail-on-finding |
Exit 1 when a result is potentially_affected or unsupported_runtime (see exit codes). |
--tool-version |
Print the tool name and version. |
Existing reports are protected unless --force is explicit. --build-source, --brand and the two observation options are checked before any input is read.
An inventory file holds one object or an array of objects. Each record requires an integer schema_version of 1 and a string node_version. Optional fields are runtime (only "node" is accepted), label, build_source and the two observation keys, each a boolean or null. Unknown keys, duplicate keys, non-finite numbers and invalid types are rejected. See examples/inventory.example.json:
[
{
"schema_version": 1,
"label": "example-service",
"runtime": "node",
"node_version": "22.23.2",
"build_source": "unknown",
"http2_service": null,
"untrusted_http2_clients": null
}
]A CSV list needs a header row with a node_version column. Optional columns are label, build_source, http2_service and untrusted_http2_clients; observation cells hold exactly true, false or nothing. Duplicate or undocumented columns, rows with a different number of fields than the header, bare quotes in unquoted cells and NUL characters are rejected. Entries without a label are named after the file and row.
| Observation | Meaning |
|---|---|
http2_service |
The operator reports an HTTP/2 service. |
untrusted_http2_clients |
The operator reports untrusted HTTP/2 clients. |
Observations are operator-supplied and unverified. They are recorded as reported_true, reported_false or not_reported and never change the version result or establish reachability.
Rules as of 2026-10-24, reviewed against the official release advisory and the linked fixed releases.
| Node.js major | Fixed version |
|---|---|
| 22.x | 22.23.2 |
| 24.x | 24.18.1 |
For each listed major, versions below its fixed version require remediation. Unlisted, custom/backported and prerelease builds need explicit evidence; do not label unknown builds safe. The vendor does not specify an introduced version in the release advisory, so earlier versions of each named major are treated as potentially_affected.
A vendor, distribution or custom build source is checked first and always yields unknown; the remaining rows apply to upstream build sources.
| Reported version | status |
version_status |
|---|---|---|
| Plain stable release below the fixed version of a listed major | potentially_affected |
below_fixed_version |
| Listed fix version, or a later plain version on that same major | fixed_version_reported |
at_or_above_fixed_version |
| Major past end of life in the embedded Node.js release schedule (0, 4–21, 23 and 25) | unsupported_runtime |
unsupported_release_line |
| Other major not listed above, including active and future majors | unknown |
release_line_not_in_rules |
| Vendor, distribution, custom or unrecognized build source | unknown |
vendor_build_unverified |
| Prerelease, nightly/rc, package-suffixed or other qualified version | unknown |
qualified_version_unverified |
Malformed or partial version such as 22.23 |
unknown |
unrecognized_or_partial_version |
| Missing or empty version | unknown |
version_not_disclosed |
unsupported_runtime is upgrade guidance, not a claim that this particular flaw exists in every historical build. A vendor, backported or custom build needs vendor evidence and stays unknown even on an end-of-life major, because its upstream version does not show which patches it contains. Version metadata is not a patch or build signature.
status |
Meaning |
|---|---|
potentially_affected |
The reported version is below the fixed release of its major and needs remediation review. |
fixed_version_reported |
The reported version is at or above the fixed release. This is not deployed-patch attestation; confirm that the running process or container actually uses that release. |
unsupported_runtime |
The reported major is past end of life. Treat it as upgrade guidance. |
unknown |
Evidence is missing or cannot be classified. Collect the exact upstream version and build provenance, then run the checker again. |
Every result has exploitability: not_tested. evidence_type is always operator_reported_version; evidence_confidence is low for classified results and insufficient for unknown. Each row also carries a recommendation and notes that repeat the advisory caveats. Operator-supplied preconditions never change the version result or establish reachability.
A reported fixed version is not a verified deployed patch, and potentially_affected is not confirmed exploitability. An operational failure (exit code 2) is not an assessment result and says nothing about the versions that were meant to be assessed.
All formats are rendered in full in memory before anything is written and carry the same evidence.
- JSON is an object with
tool,tool_version,report_schema_version(1.0),cve,rules_as_of,sources,complete,generated_atand aresultsarray. - CSV quotes every field and neutralizes formula prefixes by prepending an apostrophe to values that start with a tab, CR or LF, or with
=,+,-or@after any leading whitespace. Result rows userecord_type=result. When no result exists, arecord_type=summaryrow preservesreport_completeand the rule metadata without an asset or verdict. Notes and sources are joined with|. CSV rejects NUL characters before writing a report. - HTML is a self-contained, script-free page with a restrictive content security policy. It lists results with their evidence and limitations, the upstream fixed releases and a reminder to verify the deployed binary.
--brandsets the organization name.
CSV appends tool, tool_version and UTC generated_at to the result columns, and report_complete records whether the run finished. All result rows and empty summaries carry this report metadata. One timestamp is captured per CLI run and reused if interruption requires recovery; it is report-generation time, not evidence-acquisition time. JSON and HTML use the same timestamp.
The distributed standalone file is identified by SHA256SUMS. No Git or network lookup is needed at runtime.
- Regular UTF-8 files only.
- 4 MiB per file.
- 100 input files per run, counting
--inventoryand--csv-listtogether. - 10,000 entries per run, counting
--versionvalues and file entries together. - 512 characters per text field.
Use stable local copies of inputs. File identity checks are not a sandbox against concurrent adversarial mutation.
Reports are buffered and published atomically to files. An existing output is refused unless --force is given; without --force, publication needs hard-link support on the target filesystem and fails with a clear message otherwise. The output directory must already exist, and --output must differ from every input file even with --force.
| Code | Meaning |
|---|---|
| 0 | The run completed and the report was written. |
| 1 | --fail-on-finding was given and at least one result is potentially_affected or unsupported_runtime. The report is still written. |
| 2 | Usage, input, output or cleanup failure (an operational error, not an assessment result). This does not establish whether report bytes were already written; inspect the output before retrying. |
| 130 | Interrupted. |
- Before the output path is validated, nothing is written.
- If an interruption arrives after a complete report was published, that report is preserved and the run exits 130 without retrying.
- If an interrupt leaves file publication indeterminate, the checker exits 130 without retrying. Inspect the output before retrying; the diagnostic does not establish whether publication succeeded.
- An interruption during stdout output can leave a partial document and never appends a second report.
- In all other cases the completed results are retained and written with
complete(JSON) orreport_complete(CSV) set tofalse, and the HTML report shows an interrupted-run notice. Assess the remaining inputs again.
Keep real inventories, targets, reports and secrets outside version control. Synthetic examples are in examples/ and docs/example-report.html.
- https://nodejs.org/en/blog/vulnerability/july-2026-security-releases
- https://github.com/nodejs/security-wg/blob/main/vuln/core/184.json
- https://nodejs.org/en/blog/release/v22.23.2
- https://nodejs.org/en/blog/release/v24.18.1
End-of-life classification uses a snapshot of the Node.js release schedule reviewed on 2026-10-24 and embedded in the executable; nothing is fetched at runtime. The schedule lists the 0.x series only as 0.8, 0.10 and 0.12, so treating major 0 as end of life is an inference.
The embedded version engine is copied from node-cve-2026-105481-check commit de3161c4a8a7f7ca5c9e531df43b94e2805b52fd; no companion repository is required at runtime.
Read CONTRIBUTING.md and AGENTS.md before changing the project. This repository covers only CVE-2026-56846. Use synthetic fixtures, do not contact real targets in tests, and run python -m unittest -v, python node_cve_2026_56846_check.py --help and git diff --check before submitting. Verify current primary sources before changing rules, and keep unknown evidence uncertain.
Report flaws in this project as described in SECURITY.md. For a vulnerability in Node.js itself, follow the vendor's published security-reporting process.
Available under the GNU AGPL-3.0 (LICENSE) or a separate commercial license from CompleteTech LLC; see LICENSING.md. CompleteTech LLC names, logos and the artwork in docs/images/ are brand assets that are not covered by the code license; see BRAND_ASSETS.md. The synthetic report previews use example data and describe no real system.
