fix(privacy): drop nvidia-bug-report / dmidecode false positives

- Allowlist nvidia.com, mellanox.com, gnu.org, debian.org, ubuntu.com; the
  common "not specified / not available" FRU placeholders.
- Reject matches with no alphanumeric or <2 chars (stray ":" from a dumped
  resolv line), fru_location values that echo the field name ("Base Board
  Asset Tag", "P1-DIMMA1_AssetTag"), IPv4 embedded in a version string
  ("18:6.1.4.5"), and e-mail on kernel ring-buffer lines (driver copyright).
- Customer guess: don't report a single-hit low-confidence domain at all -
  "unidentified" beats guessing nvidia.com from a driver comment.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
This commit is contained in:
Mikhail Chusavitin
2026-09-02 16:27:17 +03:00
co-authored by Claude Sonnet 5
parent c9748f3830
commit 551d8e450d
5 changed files with 105 additions and 12 deletions
+7
View File
@@ -47,6 +47,13 @@ via `parser.PrivacyScanEnabled()`.
a `resolv` finding, not also a `domain` one).
- A match that parses as an IP but is not sensitive (see IP policy) is dropped
regardless of the rule that produced it.
- Noise guards: matches shorter than 2 chars or with no alphanumeric are
dropped; `fru_location` values that echo the field name (`Base Board Asset
Tag`, `P1-DIMMA1_AssetTag`) are dropped; IPv4 inside a longer dotted-number
run (`18:6.1.4.5`) is not an address; e-mail on a kernel ring-buffer line
(`[ 8.07] ...`) is driver/copyright text, not customer contact.
- Customer guess: a single-hit low-confidence domain is not reported at all
(better "unidentified" than guessing `nvidia.com` from a driver comment).
## Rule catalogue (`rules.go`)