Offline SonicWall SMA1000 evidence checker for CVE-2026-102255. Reported build and configuration evidence. Standalone Python; JSON/CSV/HTML reports; no exploit payloads.
SonicWall SMA1000 (6210, 7210, 8200v): Server-side request forgery
CVE-2026-102255 · offline, version-only checker
sonicwall_sma1000_cve_2026_102255_check.py is a standalone, standard-library-only
Python 3.10+ tool. It triages reported SonicWall SMA 1000 (6210, 7210, 8200v)
build strings against the vendor's SNWLID-2026-0017 rule table for
CVE-2026-102255, a server-side request forgery (SSRF) issue.
It is offline and version-only: it makes no network requests, never contacts or
probes a device, and has no network mode, because a version string cannot be answered
by passive metadata. Exploitability is always not_tested.
[!IMPORTANT] Version claims are evidence, not verified runtime state.
potentially_affectedis not confirmed exploitability,fixed_version_reportedis not a verified deployed patch, andunknownis never a clean result.
| Field | Scope |
|---|---|
| Exploit identifier | CVE-2026-102255 |
| Aliases for this same exploit | None recorded |
| Sources reviewed | 2026-10-24 |
| Implementation state | implemented (offline version triage) |
| Network mode | None |
This repository is dedicated to CVE-2026-102255. Aliases and multiple affected products identify this same exploit; they do not create additional scanner scopes. Future fixes for different vulnerabilities belong in independent repositories.
Python 3.10 or later and nothing else: no third-party dependencies and no install step. CI runs the tests on Ubuntu, Windows and macOS with Python 3.10 and 3.14.
# Triage one reported build (prints a JSON report)
python sonicwall_sma1000_cve_2026_102255_check.py --version 12.4.3-03526
# Triage the synthetic example inventory into a standalone HTML report
python sonicwall_sma1000_cve_2026_102255_check.py --inventory examples/inventory.example.json --format html --output report.html
# Show all options and exit-status meanings
python sonicwall_sma1000_cve_2026_102255_check.py --help
At least one of --version, --inventory or --csv-list is required; otherwise the
run exits with status 2.
| Input | Flags |
|---|---|
| Reported build(s) | --version BUILD (repeatable), with optional --branch, --model (6210, 7210 or 8200v) and --workplace-interface-access true|false|unknown; these three apply only to --version values |
| JSON inventory | --inventory FILE (repeatable): one object or an array of objects |
| CSV list | --csv-list FILE (repeatable): header row with a version column |
Inputs can be combined in one run. Builds must match
<major>.<minor>.<patch>-<build> (for example 12.4.3-03526); anything else is
classified unknown.
| Entry field | JSON inventory | CSV list |
|---|---|---|
schema_version | Required, exactly the integer 1 | Not allowed as a column |
version | Optional (missing is unknown) | Required column |
id (or label) | Optional | Optional |
branch, model | Optional | Optional |
workplace_interface_access | Boolean or null | Optional |
custom_build | Boolean | Optional |
version_candidates | List of strings | Optional, semicolon separated |
Input is validated strictly. Unknown fields or columns, wrong types, duplicate keys,
malformed CSV, oversized values, symlinks and non-regular files are operational
errors (exit 2), not assessment results. Limits: 1 MiB per input file, 10,000
entries per run, 128 characters per field and at most 50 version_candidates per
entry. docs/REFERENCE.md lists the full input rules,
including accepted boolean spellings and whitespace handling, and
examples/inventory.example.json is a working
inventory.
| Option | Behavior |
|---|---|
--format json|csv|html | Report format; json is the default |
--output FILE | Write the report to FILE instead of stdout |
--force | Allow --output to overwrite an existing file |
--output is validated before any input is read: it may not be an input file and
is not overwritten without --force. The report is written to a temporary file in
the same directory and published atomically; without --force, publication cannot
replace a destination that appears mid-run, and the run exits 2 where hard links
are unsupported. Report text that cannot be encoded as UTF-8 also exits 2.
| Code | Meaning |
|---|---|
0 | Completed; no potentially_affected result. unknown results are not a clean result. |
1 | Completed; at least one potentially_affected result, or any unknown result with --strict. |
2 | Operational or usage error (bad input, unreadable file, refused overwrite). Not an assessment result. |
130 | Interrupted (Ctrl-C). See below. |
--strict makes any unknown result count as a finding (exit 1), which suits
automation that must not treat unclassified builds as clean.
The report carries a complete flag (and a CSV complete column). If the run is
interrupted after the output path is validated but before publication, the results
completed so far are written with complete: false and the notice says the run was
interrupted; the exit status is 130.
130 without claiming a report was published. Check the report artifact
before running again.docs/REFERENCE.md describes the reconciliation details.