feat(privacy): detect an already-sanitized source

Scan now reports PrivacyScan.Sanitized {detected, markers, strong, files,
evidence}. SanitizationMarkers recognises a value slot filled with one
repeated placeholder + separators (xxxxx.xxxx.xx, x@xxxx.xxxx.xx,
000.00.00.0, a decoy timezone) - it matches the shape, not the literal "x",
so evolving the redaction mechanism still trips it.

detected requires corroboration: strong>=2, or strong>=1 && markers>=3, or
markers>=4. A single filler-looking token is reported (markers:1) but never
asserted as sanitized, so a partial future pass or a coincidence does not
read as "done". 0.0.0.0 / 000 / UTC / Etc/UTC are too plausibly intentional
and do not count.

UI: the Customer-data panel shows "файл уже обезличен" and hides the
sanitize button when detected.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_014jDYM1nnoZZ3vFz23DDaV1

bible-local/03-api.md

@@ -155,9 +155,11 @@ Returns normalized parse and collection issues combined from:
 ### `GET /api/privacy-scan`
 
 Returns the customer-data scan for the current dataset (`models.PrivacyScan`:
-`files_scanned`, `customers[]`, `findings[]`, `summary`).
-Returns `{ "loaded": false }` when nothing is loaded or the scan produced
-nothing. Detection only; see `docs/privacy-scan.md`.
+`files_scanned`, `customers[]`, `findings[]`, `summary`), plus `sanitizable`
+(bool) and, when the source already looks de-identified, `sanitized`
+(`{detected, markers, strong, files, evidence}`).
+Returns `{ "loaded": false }` when nothing is loaded. Detection only; see
+`docs/privacy-scan.md`.
 
 ### `POST /api/sanitize`
 

bible-local/10-decisions.md

@@ -1974,6 +1974,10 @@ with **same-length neutral fillers**, in place.
   reported, not edited.
 - Re-parsing a sanitized dump yields the identical hardware inventory
   (verified on Dell TSR, xFusion, Inspur onekeylog, H3C `.sds`).
+- The scan detects an already-sanitized source
+  (`PrivacyScan.Sanitized`, `privacy/sanitized.go`): it matches the *shape*
+  of a filled value slot, not the literal `x`, and needs corroborating
+  markers so a partial future redaction pass does not read as "done".
 - Full contract and rule list: `bible-local/docs/log-sanitization.md`.
 - Tests: `internal/sanitize` (`TestRedactText_*`, `TestSanitize_Tar*`,
   `TestSanitize_Zip_*`, `TestSanitize_BinaryMemberFlagged`,

bible-local/docs/log-sanitization.md

@@ -21,6 +21,11 @@ scan (see `privacy-scan.md`). ADL-067.
 - **No marker.** Nothing is stamped into the file.
 - **Idempotent.** The fillers are on the privacy allowlist (`isRedactionFiller`),
   so a re-scan finds nothing and a second `Sanitize` is a no-op.
+- **Detectable.** The privacy scan reports an already-sanitized file in
+  `PrivacyScan.Sanitized` - see `privacy-scan.md`. It needs corroborating
+  markers, so a single filler-looking token (or a partial future redaction
+  pass) does not read as "done", and the UI hides the sanitize button when a
+  file already looks de-identified.
 
 ## Entry point
 

bible-local/docs/privacy-scan.md

@@ -115,6 +115,23 @@ default values (`Asia/Shanghai`, `To Be Filled By O.E.M.`, `NULL`, `0.0.0.0`).
 neutral fillers (see `log-sanitization.md`). The fillers are recognised by
 `isRedactionFiller` so a scan of a sanitized file is clean.
 
+## Already-sanitized detection (`sanitized.go`)
+
+`Scan` also reports whether the source already looks de-identified, in
+`PrivacyScan.Sanitized` (`{detected, markers, strong, files, evidence}`).
+
+`SanitizationMarkers(line)` recognises a **value slot filled with one repeated
+placeholder + separators** - `xxxxx.xxxx.xx`, `x@xxxx.xxxx.xx`, `000.00.00.0`,
+a decoy timezone (`Antarctica/McMurdo`, `Etc/Universal`, ...). It matches the
+*shape*, not the literal `x`, so changing the filler character later still
+trips it. `0.0.0.0`, `000`, `UTC`, `Etc/UTC` are too plausibly intentional and
+do not count.
+
+`detected` requires corroboration - `strong >= 2`, or `strong >= 1 && markers
+>= 3`, or `markers >= 4`. A single filler-looking token is reported
+(`markers: 1`) but never asserted as sanitized, so a partial future redaction
+pass or a coincidence does not read as "done".
+
 ## Customer guess (`customer.go`)
 
 Findings in `domain`, `resolv`, `ad_ldap`, `cert`, `nsupdate`,