Skip to content

About

Offline Node.js evidence checker for CVE-2026-58045: zlib synchronous API spoofed TypedArray crash. Single-file Python; JSON/CSV/HTML reports; no exploit payloads.

Topics

Resources

Code of conduct

Contributing

Security policy

Stars

0 stars

Watchers

0 watching

Forks

node-cve-2026-58045-check by CompleteTech LLC

Tests Python 3.10+ No runtime dependencies AGPL-3.0 or commercial license

node-cve-2026-58045-check

Offline version-evidence checker for CVE-2026-58045 — Node.js: zlib synchronous API spoofed TypedArray crash.

It compares Node.js runtime versions that you report against the fixed releases named in the Node.js security advisory, and writes the result as CSV, JSON or a self-contained HTML report. Python 3.10+, standard library only; the executable is a single file and can be copied on its own.

Contents

Scope and honest limits

This repository covers only CVE-2026-58045. The checker is triage evidence, not a runtime assessment.

  • Offline. It never contacts websites, executes Node.js or project code, or sends exploit payloads. No network mode is implemented, so there is no TLS, robots or pacing behavior to configure.
  • No discovery. It does not scan the current machine or any target. A public URL is not an inventory. Supply an exact version obtained from the deployed process, and only for systems you are authorized to inventory.
  • Uncertainty stays visible. Hidden versions, custom builds, conflicting evidence, and uncovered branches remain unknown.
  • Version evidence only. A reported fixed version is not proof of a deployed patch, and a version below the fix is not proof of exploitability. Every result carries exploitability: not_tested.

Requirements

  • Python 3.10 or later; no third-party packages.
  • Exact Node.js version strings from the running process or container (for example, the output of node --version on that runtime, with a leading v accepted).
  • Optionally, the build source (upstream or a vendor name) and operator-reported precondition observations. See Inventory formats.

CI runs the test suite and --help on Python 3.10 and 3.14, on Ubuntu and Windows.

Quick start

# Assess one reported version and print JSON to stdout
python node_cve_2026_58045_check.py --version 22.23.2 --format json

# Assess the synthetic example inventory and save an HTML report
python node_cve_2026_58045_check.py --inventory examples/inventory.example.json --format html --output report.html

# Assess your own CSV list (CSV to stdout by default); exit 1 on a finding
python node_cve_2026_58045_check.py --csv-list inventory.csv --fail-on-finding

# Show all options, then run the local tests
python node_cve_2026_58045_check.py --help
python -m unittest -v

Synthetic report preview

Node.js synthetic report

Open the synthetic example report. Example data only; this is not a live assessment.

Usage

Provide at least one of --version, --inventory or --csv-list; otherwise the checker reports a usage error. The three sources can be combined in one run, and every input becomes one result row. JSON, CSV and HTML are rendered from the same evidence.

Option Description
--version VERSION (alias --node-version) Assess an explicitly supplied Node.js version. Repeatable; each value describes a separate runtime. A leading v is accepted.
--build-source SOURCE Provenance of the --version builds, at most 512 characters. unknown (the default), upstream, official or official_upstream are compared as upstream; any other value, such as a vendor name, yields unknown with vendor_build_unverified.
--zlib-synchronous-api {true,false} Operator-reported observation applied to the --version inputs.
--attacker-controlled-typedarray-object {true,false} Operator-reported observation applied to the --version inputs.
--inventory FILE JSON object or array of entries. Repeatable.
--csv-list FILE CSV with a node_version column. Repeatable.
--format {csv,json,html} Report format. Default: csv.
--output FILE Save the report to a new file. Default: stdout.
--force Replace an existing output file. Without it, an existing file is never overwritten.
--brand NAME Organization name shown in the HTML report, at most 512 characters. Default: CompleteTech LLC.
--fail-on-finding Exit 1 when any result is potentially_affected or unsupported_runtime. See exit codes.
--tool-version Print the tool name and version.

Output protections:

  • --output must name a file in an existing directory. An existing report is protected unless --force is explicit, and an input file cannot be overwritten even with --force.
  • Reports are rendered fully in memory, then published atomically to files. Without --force the checker publishes with a hard link and fails with an error rather than overwrite if the filesystem cannot do that.
  • The precondition flags and --build-source apply only to --version inputs; inventory and CSV entries carry their own values.

Inventory formats

Inventories are optional; --version needs none. Inventory entries may include operator-reported precondition observations. These never change the version result or establish reachability.

Observation key Meaning
zlib_synchronous_api Operator-reported: zlib synchronous API
attacker_controlled_typedarray_object Operator-reported: attacker-controlled TypedArray object

Each observation is true, false or unreported, and appears in results as reported_true, reported_false or not_reported.

JSON (--inventory)

The file holds one object or an array of 1 to 10,000 objects. Each record requires integer schema_version: 1 and string node_version. Optional fields are runtime (only "node"), label, build_source, and the observation keys above, each boolean or null. Unknown keys, duplicate keys, non-finite numbers and invalid types are rejected. examples/inventory.example.json is a synthetic example:

[
  {
    "schema_version": 1,
    "label": "example-service",
    "runtime": "node",
    "node_version": "22.23.2",
    "build_source": "unknown",
    "zlib_synchronous_api": null,
    "attacker_controlled_typedarray_object": null
  }
]

CSV (--csv-list)

The first row must be a header containing node_version. Optional columns are label, build_source, zlib_synchronous_api and attacker_controlled_typedarray_object; observation cells are exactly true, false, or blank. schema_version and runtime are not CSV columns. Duplicate or undocumented columns, rows whose field count differs from the header, bare quotes inside unquoted cells, text after a closing quote, NUL characters and files with no data rows are rejected.

node_version,label,build_source,zlib_synchronous_api
22.23.2,example-service,upstream,

When label is absent (JSON or CSV), results are labeled <file name without extension>:<entry number>, counting from 1.

Fixed versions and classification

Reviewed 2026-10-24 against the official release advisory and linked fixed releases.

Node.js major Fixed version
22.x 22.23.2
24.x 24.18.1
26.x 26.5.1

The vendor does not specify an introduced version in the release advisory. 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.

How a reported version is classified, in the order the checks apply:

Reported evidence status version_status
No version, or only whitespace unknown version_not_disclosed
Valid X.Y.Z followed by a suffix (prerelease, nightly/rc, package suffix, custom build) unknown qualified_version_unverified
Anything else that is not a plain X.Y.Z release unknown unrecognized_or_partial_version
Plain version, but a vendor, distribution, custom or unrecognized build source unknown vendor_build_unverified
Plain version on a listed major, below its fixed version potentially_affected below_fixed_version
Plain version on a listed major, at or above its fixed version fixed_version_reported at_or_above_fixed_version
Unlisted major that is end-of-life (0, 4–21, 23, 25) unsupported_runtime unsupported_release_line
Any other unlisted major, including future majors unknown release_line_not_in_rules

Notes:

  • Whitespace is trimmed and at most one leading v is removed before comparison. If the engine is given conflicting versions for one runtime, the result also stays unknown (conflicting_versions).
  • Earlier versions of each named major are potentially_affected because no introduced version is established. Listed fix versions and later plain versions on the same major are fixed_version_reported.
  • Old end-of-life majors produce unsupported_runtime as upgrade guidance, not a claim that this particular flaw exists in every historical build. End-of-life status comes from the Node.js release schedule as reviewed on 2026-10-24 (see Sources and provenance); major 0 is an inference because the schedule lists only 0.8, 0.10 and 0.12.
  • Vendor, backported and custom builds require vendor evidence and remain unknown. The upstream version does not show which patches a vendor build contains.

Interpreting results

status Meaning
potentially_affected The reported version is below the fixed release of its major and needs remediation review. This is not confirmed exploitability.
fixed_version_reported The reported version is at or above the fixed release. This is not deployed-patch attestation: confirm the running process or container actually uses that release.
unsupported_runtime The reported major is end-of-life. Treat it as upgrade guidance, not a finding about this specific flaw.
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. Nothing in this tool tests, reproduces or confirms exploitation or a production deployment.
  • evidence_type is operator_reported_version. evidence_confidence is low for classified results and insufficient for unknown.
  • Operator-supplied precondition observations are unverified. They never change status or version_status and do not establish reachability.
  • version_status gives the specific reason for each status; see the classification table.
  • Each result includes a recommendation and notes, which carry the advisory caveats alongside the per-input explanation.
  • Operational failures (unreadable or invalid input, unwritable output, usage errors) are not assessment results. They exit with status 2 and an error message on stderr.

Report output

Choose the format with --format; the default is CSV on stdout.

Field Value
tool, tool_version The canonical tool name and version, in every format. --tool-version prints them.
generated_at UTC ISO-8601 timestamp. One is captured per CLI run and reused if an incomplete report is written after an interrupt.
rules_as_of The advisory review date (2026-10-24), not the time the evidence was acquired.
report_schema_version Report schema version (1.0).
cve, sources The CVE identifier and primary source URLs.
  • JSON is one object with tool, tool_version, report_schema_version, cve, rules_as_of, sources, complete, generated_at and a results array. Each result holds the input label and source, reported version and build source, status, version_status, fixed_version, exploitability, evidence type and confidence, the precondition observations, recommendation and notes.
  • CSV quotes every field and neutralizes formula prefixes (=, +, -, @) after leading whitespace. List values are joined with |. Provenance columns (tool, tool_version, generated_at) follow the existing fields. record_type=result marks each completed assessment, and report_complete records whether the run finished. When no result exists, such as after an early interrupt, a single record_type=summary row preserves report_complete, rule metadata and provenance; no asset, verdict or assessment is invented. NUL characters are rejected before any report is written.
  • HTML is a self-contained, script-free page with a restrictive content security policy and escaped values. It shows the report schema, each result's evidence type and confidence, per-input notes, and the table of fixed releases, and it sorts results by severity. --brand changes only the organization name in this report.

Limits, exit codes and interruption

Input limits:

  • Regular UTF-8 files only (a UTF-8 byte-order mark is accepted).
  • 4 MiB per file, 100 files per run (--inventory and --csv-list combined), 10,000 entries per run (including --version values), and 512 characters per text field.
  • Use stable local copies. File identity checks are not a sandbox against concurrent adversarial mutation.
Exit code Meaning
0 The run completed and the report was written. Findings do not change this unless --fail-on-finding is set.
1 Only with --fail-on-finding: at least one result is potentially_affected or unsupported_runtime.
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:

  • Before the output path is validated, nothing is written.
  • During assessment, completed results are written with complete: false (JSON) or report_complete=false (CSV), or an "Interrupted run" notice (HTML). Remaining inputs are not assessed; run them again. This happens when output is possible, to --output or stdout.
  • An interrupt after a complete report is published preserves that report and 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 interrupt while writing to stdout can leave a partial document, and the checker never appends a second report. A second interrupt during recovery may leave no partial report.

Handling inventories and reports

Keep real inventories, targets, reports and secrets outside version control. Reports contain whatever labels and versions you supplied. Synthetic examples live in examples/ and docs/example-report.html.

Sources and provenance

Primary sources, reviewed 2026-10-24:

End-of-life majors come from the Node.js release schedule, reviewed 2026-10-24 and embedded in the checker as constants. It is not fetched at run time.

The embedded version engine is copied from node-cve-2026-105481-check commit de3161c4a8a7f7ca5c9e531df43b94e2805b52fd; no companion repository is required at runtime. Version metadata is not a patch or build signature.

Project documents

Document Contents
docs/REFERENCE.md CLI and evidence reference
docs/IMPLEMENTATION.md Implementation and verification status
docs/example-report.html Synthetic example report
advisory.json Source-preserving advisory metadata for the offline rule
AGENTS.md Repository instructions for contributors and automation

Contributing and security

Open a focused issue and pull request, use synthetic fixtures, and never contact real targets in tests. Before submitting, run:

python -m unittest -v
python node_cve_2026_58045_check.py --help
git diff --check

Verify current primary sources before changing rules, keep unknown evidence uncertain and vendor provenance separate, and keep exploitability at not_tested. Hosted CI must pass for the exact reviewed head. See CONTRIBUTING.md, including licensing of contributions.

To report a flaw in this project, follow SECURITY.md. For a vulnerability in Node.js itself, use the Node.js project's published reporting process.

License and brand assets

Available under GNU AGPL-3.0 or a separate commercial license from CompleteTech LLC. See LICENSE and LICENSING.md.

CompleteTech LLC names, logos and brand assets, including the images in docs/images/, are not licensed under the AGPL-3.0 or a commercial code license. See BRAND_ASSETS.md. The synthetic report previews use example data and describe no real system.

Copyright © 2026 CompleteTech LLC.

About

Offline Node.js evidence checker for CVE-2026-58045: zlib synchronous API spoofed TypedArray crash. Single-file Python; JSON/CSV/HTML reports; no exploit payloads.

Topics

Resources

Code of conduct

Contributing

Security policy

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages