Offline evidence checker for CVE-2026-105482 — Node.js: DNS resolution NUL-name truncation. It compares Node.js runtime versions that you report against the fixed releases named in the Node.js October 2026 security advisory, and writes the result as CSV, JSON or HTML.
Python 3.10+, standard library only. The single executable, node_cve_2026_105482_check.py, can be copied on its own.
Version evidence, not an assessment of a live system. A result describes the version you supplied. It does not show that a patch is deployed, that the flaw is reachable, or that it can be exploited.
- Scope and limits
- Requirements
- Quick start
- Report preview
- Fixed versions and rule snapshot
- Inputs
- Output
- Interpreting results
- Exit codes and interruption
- Input limits and file handling
- Sources and provenance
- Development and documentation
- Security, licensing and brand
This repository covers only CVE-2026-105482. It is a small, offline triage tool.
- It does: compare operator-reported runtime versions with the vendor's fixed releases and report the outcome with the evidence limits that apply.
- It does not: scan the current machine, contact websites or any network service, execute Node.js or project code, or send exploit payloads. A public URL is not an inventory.
- Unknown stays unknown. Hidden versions, custom or vendor builds, prereleases, conflicting evidence and uncovered release lines are reported as
unknown; they are never labelled safe. - Exploitability is never tested. Every result carries
exploitability: not_tested. - A reported fixed version is not a verified patch.
fixed_version_reportedmeans the supplied version is at or above the vendor's fixed release. It is not attestation of what is deployed.
Supply data only for systems you are authorized to assess.
- Python 3.10 or later. The CI workflow runs Python 3.10 and 3.14 on Ubuntu and Windows.
- No third-party packages and no network access.
- An exact Node.js version for each runtime, obtained from the deployed process or container (for example,
node --versionrun where the application runs), plus its build source if known. The checker never obtains this itself.
# Assess one version and print JSON to stdout
python node_cve_2026_105482_check.py --version 22.23.4 --format json
# Assess the synthetic example inventory and save an HTML report
python node_cve_2026_105482_check.py --inventory examples/inventory.example.json --format html --output report.html
# Assess a CSV list and exit 1 if any result is a finding
python node_cve_2026_105482_check.py --csv-list inventory.csv --fail-on-finding
# Show all options, or run the local synthetic tests
python node_cve_2026_105482_check.py --help
python -m unittest -vinventory.csv above is your own file; see CSV list for its columns. With no --format, the report is CSV. With no --output, it goes to stdout.
The preview shows the HTML report in light and dark themes. Open the synthetic example report. It uses example data only and is not a live assessment.
The rules are dated 2026-10-24 (rules_as_of in reports). They were reviewed against the official Node.js release advisory, published 2026-10-21, and the linked fixed releases.
| Node.js major | Fixed version |
|---|---|
| 22.x | 22.23.4 |
| 24.x | 24.22.1 |
| 26.x | 26.12.1 |
How a plain stable version such as 22.20.0 or v24.22.1 is classified:
- Below the fixed version of a listed major:
potentially_affected. The vendor does not establish an introduced version, so every earlier version of each named major is treated this way, and versions below the fixed version require remediation. - At or above the fixed version on the same major:
fixed_version_reported. - A major that is neither listed nor end-of-life (for example, a newer major, or 1–3):
unknown. It is not assumed to be patched. - An end-of-life major:
unsupported_runtime. This is upgrade guidance, not a claim that this flaw exists in every historical build. The vendor states that end-of-life versions are always affected by a security release; the checker does not assert that for this specific flaw. - Prereleases, nightly or rc builds, package suffixes, and vendor, backported or custom builds:
unknown. They need explicit vendor evidence; do not label unknown builds safe.
End-of-life majors are 0 (an inference, because the schedule lists only 0.8, 0.10 and 0.12), 4–21, 23 and 25, taken from the Node.js release schedule as reviewed on 2026-10-24. This list is built into the checker; the schedule is not fetched at run time.
A version number is not a patch or build signature. Vendor, backported and custom builds may contain fixes that their version number does not show, or may lack them.
At least one of --version, --inventory or --csv-list is required. They can be combined; results appear in the order --version values, then --inventory files, then --csv-list files. Each version or inventory row describes a separate runtime.
| Option | Purpose |
|---|---|
--version VERSION (alias --node-version) |
Assess an exact Node.js version. A leading v is accepted. Repeatable; each value is a separate runtime. |
--build-source SOURCE |
Provenance of the --version builds, at most 512 characters. See Build source. |
--dns-resolve-or-reverse {true,false}--untrusted-hostnames {true,false}--hostname-validation-before-resolution {true,false} |
Operator observations applied to every --version input. See Precondition observations. |
--inventory FILE |
JSON inventory file. Repeatable. |
--csv-list FILE |
CSV list file. Repeatable. |
--format {csv,json,html} |
Report format. Default: csv. |
--output FILE |
Save the report to a new file. Default: stdout. |
--force |
Allow --output to replace an existing file. |
--brand NAME |
Organization name shown in the HTML report, at most 512 characters. Default: CompleteTech LLC. It does not affect CSV or JSON. |
--fail-on-finding |
Exit 1 when any result is potentially_affected or unsupported_runtime. See Exit codes and interruption. |
--tool-version |
Print the tool name and version. |
--help |
Print usage and exit-status help. |
--build-source and the build_source inventory field state where a build came from. An empty value, unknown (the default), upstream, official or official_upstream is compared as an official upstream build; matching ignores case and surrounding whitespace. Any other value, such as a distribution or vendor name, yields unknown with version_status vendor_build_unverified, because the upstream version does not show which patches or backports that build contains. This check comes before the version comparison, so it applies to end-of-life and unlisted majors as well.
Because the default is compared as upstream, the answer is only as reliable as the claim that the build is an unmodified official release.
These optional, operator-supplied observations are recorded next to the result. They are unverified, never change the version result, and do not establish reachability.
| Inventory key | CLI option | Observation |
|---|---|---|
dns_resolve_or_reverse |
--dns-resolve-or-reverse |
DNS resolve or reverse |
untrusted_hostnames |
--untrusted-hostnames |
Untrusted hostnames |
hostname_validation_before_resolution |
--hostname-validation-before-resolution |
Hostname validation before resolution |
Reports show each observation as reported_true, reported_false or not_reported.
A JSON inventory is one object or an array of objects, with 1 to 10,000 entries.
| Field | Required | Type |
|---|---|---|
schema_version |
Yes | Integer 1 |
node_version |
Yes | String |
runtime |
No | "node" (the only accepted value) |
label |
No | String; defaults to <file name without extension>:<entry number> |
build_source |
No | String; see Build source |
| Observation keys above | No | Boolean or null |
Unknown keys, duplicate keys, non-finite numbers, and invalid types are rejected. See examples/inventory.example.json for a synthetic example.
A CSV list needs a header row with a node_version column. Optional columns are label, build_source, and the three observation keys; schema_version and runtime are not CSV columns. Observation cells must be exactly true or false, or blank. Undocumented or duplicate columns, rows with a different number of fields than the header, quotes inside unquoted cells, text after a closing quote, unterminated quotes, and NUL characters are rejected.
Reports are CSV (default), JSON or HTML. All three are built from the same evidence. The report is rendered in full before anything is written.
- Without
--output, the report is written to stdout. - With
--output, the directory must already exist and the path must be a new regular file. Reports are buffered and published atomically, without overwriting. If the filesystem cannot create hard links, the run fails and asks for another filesystem or an explicit--force. - An existing file is replaced only with
--force. --outputcan never be one of the input files, even with--force.
Each result records the input label and source, the reported node_version and build_source, status, version_status, fixed_version, exploitability (always not_tested), evidence_type (operator_reported_version), evidence_confidence (insufficient for unknown results, otherwise low), cve, rules_as_of, the observation values, a recommendation, notes and sources, and report_schema_version (1.0). Every result's notes include the advisory caveats.
- JSON: one object with
tool,tool_version,report_schema_version,cve,rules_as_of,sources,complete,generated_atand aresultsarray in input order. - CSV: every field is quoted, and values that begin with a formula prefix (
=,+,-or@, even after leading whitespace) or a tab, carriage return or line feed are neutralized with a leading apostrophe. Lists such asnotesare joined with|. Rows haverecord_type=result. When no result exists (for example, an interruption before any input finished), a singlerecord_type=summaryrow keepsreport_completeand the rule metadata, with no asset or verdict.report_completeistrueorfalse. - HTML: a script-free page with a restrictive content security policy. Results are sorted by severity:
potentially_affected,unsupported_runtime,unknown, thenfixed_version_reported. An interrupted run is flagged on the page.
CSV appends tool, tool_version and generated_at to every result and to the summary row. These fields identify the checker and the timezone-aware UTC report-generation time. They are separate from rules_as_of and do not show when the version evidence was collected or whether a deployed patch is present. One run timestamp is reused for JSON, CSV, HTML and interruption recovery.
status |
Meaning |
|---|---|
potentially_affected |
The reported version is below the fixed version of its major. It needs remediation review; this is not confirmed exploitability. |
fixed_version_reported |
The reported version is at or above the fixed version. This is not deployed-patch attestation. |
unsupported_runtime |
The major is end-of-life. Treat it as upgrade guidance, not as proof that this flaw is present. |
unknown |
The evidence is missing or cannot be classified. Investigate; do not treat it as safe. |
version_status gives the specific reason:
version_status |
Used with | Cause |
|---|---|---|
below_fixed_version |
potentially_affected |
Below the fixed release of a listed major |
at_or_above_fixed_version |
fixed_version_reported |
At or above the fixed release of a listed major |
unsupported_release_line |
unsupported_runtime |
End-of-life major |
release_line_not_in_rules |
unknown |
Major not covered by this rule snapshot |
vendor_build_unverified |
unknown |
Vendor, distribution, custom or unrecognized build source |
qualified_version_unverified |
unknown |
Prerelease, nightly/rc, package suffix or other qualified version |
unrecognized_or_partial_version |
unknown |
Not a recognizable major.minor.patch version |
version_not_disclosed |
unknown |
No version was reported |
Points to keep in mind:
exploitabilityis alwaysnot_tested. Nothing in a report is evidence that the flaw was exploited or can be exploited.- Precondition observations are operator-supplied and unverified; they never change
statusorversion_status. - The
recommendationin each result is next-step guidance. Forfixed_version_reported, it is to confirm that the running process or container really uses that release, not only a patched workstation or lockfile. - A run that completes with no findings shows only that the supplied versions were not classified as findings. It does not show that the real deployment is patched.
| Code | Meaning |
|---|---|
0 |
The run completed and the report was written. This holds even when results are findings or unknown. |
1 |
Only with --fail-on-finding: at least one result is potentially_affected or unsupported_runtime. unknown results do not trigger it. |
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. |
Interruption behavior:
- Interrupted before argument and output-path validation finishes: nothing is written.
- Interrupted while inputs are loaded or assessed: the completed results are written with
complete: false(JSON) orreport_complete=false(CSV), and the HTML page flags the run as interrupted. Assess the remaining inputs again. - Interrupted after the complete report was published: that report is preserved and the run exits 130 without retrying.
- Interrupted during file publication: the outcome may be indeterminate. The checker exits 130 without retrying. The diagnostic does not say whether publication succeeded, so inspect the output before retrying.
- Interrupted while writing to stdout: the document can be partial, and a second report is never appended.
- A second interrupt while the partial report is being written exits 130, and the partial report may not have been written.
| Limit | Value |
|---|---|
Input files per run (--inventory and --csv-list combined) |
100 |
| Size per input file | 4 MiB |
| Entries per run (all sources combined) | 10,000 |
| Characters per text field | 512 |
- Inputs must be regular UTF-8 files.
- Use stable local copies of inputs and outputs. File identity checks are not a sandbox against concurrent adversarial mutation.
- Keep real inventories, targets, reports and secrets outside version control. Use only synthetic data in examples and tests.
Primary sources, as reviewed on 2026-10-24:
- Node.js October 2026 security releases
- Node.js v22.23.4 release
- Node.js v24.22.1 release
- Node.js v26.12.1 release
- Node.js release schedule, used for the end-of-life majors
The version-comparison engine is copied from node-cve-2026-105481-check commit de3161c4a8a7f7ca5c9e531df43b94e2805b52fd. No companion repository is required at runtime. Reports list the four advisory and release links in sources.
Run the local synthetic tests with python -m unittest -v. CI runs them, plus --help, on Ubuntu and Windows with Python 3.10 and 3.14. Tests do not contact real targets.
| File | Contents |
|---|---|
node_cve_2026_105482_check.py |
The checker |
SHA256SUMS |
Checksum of the executable |
advisory.json |
Source-preserving advisory metadata for the offline rule |
examples/ |
Synthetic inventory example |
docs/REFERENCE.md |
CLI and evidence reference |
docs/IMPLEMENTATION.md |
Implementation and verification status |
docs/example-report.html |
Synthetic example report |
AGENTS.md |
Repository instructions for maintainers and coding agents |
The implementation was independently reviewed against the primary sources on 2026-10-24. No deployed-target assessment, runtime attestation or exploit verification is claimed. See docs/IMPLEMENTATION.md for status.
To contribute, read CONTRIBUTING.md: open a focused issue and PR, use synthetic fixtures, verify primary sources before changing rules, and keep unknown evidence uncertain.
- Security: see SECURITY.md to report a flaw in this project privately. For a flaw in Node.js or in a service you assess, follow that vendor's or owner's reporting process. Editing this repository authorizes no scanning of real targets.
- License: GNU AGPL-3.0 (LICENSE) or a separate commercial license from CompleteTech LLC. See LICENSING.md. Primary advisory links give provenance; this repository does not relicense those external publications.
- Brand assets: CompleteTech LLC names, logos and the images in
docs/images/are not licensed under the AGPL-3.0 or a commercial code license. See BRAND_ASSETS.md.
Copyright © 2026 CompleteTech LLC.
