fix: review findings F03-F06 — quotas, rule hierarchy and a validator that checks

F03 — application records
- Quota already removed with F02's tracker; verified uncapped and that
  "interview reached" survives a later withdrawal.
- Aker BP: application date confirmed 2026-07-29 by the user, settling the
  one-day conflict. Its Adjacent-despite-79 classification is kept, NOT
  reclassified: the user does not remember the rationale and judged it
  immaterial on a closed role, and silently reclassifying to make the
  arithmetic tidy is the exact objection the review raised. Recorded as an
  explicit override reason with a rubric version instead.

F04 — rule hierarchy
- Three levels stated in AGENTS.md and CLAUDE.md: immutable evidence rules,
  standing user preferences, market/employer presentation.
- Deliverables separated from QA. The .tex stays the artefact of record; you
  still compile, into the application's own output/ folder, because the user
  cannot approve what he cannot open.
- `resume_reference.md` said "Capgemini/Bundeswehr: include only when directly
  relevant", contradicting the standing never-include rule. Split and fixed.

F05 — a green validator now means something
- Per-claim `forbidden` wording is enforced. 14 of 26 claims carried such a
  list and NONE of it was checked, which is why "owned the full ML lifecycle"
  passed. Matching tolerates inserted adjectives: the real regression read
  "Built a decentralized Data Mesh" and one word defeated the literal list.
- Added identity (e-mail/phone), employer-date, placeholder and
  document-exclusion checks. Capgemini's omission rule now has a canonical
  home in claims.json instead of living only in memory and config.md.
- Registered Groovy and Snowflake as forbidden skills. Both were absent from
  claims.json entirely, so "forbidden" existed in prose with nothing enforcing
  it.
- PASS no longer reads as a truth certificate: it prints what it did not check
  — scope implications, bullet-to-claim support, whether a number is real.

F06 — the scanner stopped fighting the documents
- LaTeX comments are no longer claims; `% never claim Terraform` was a finding
  before, which taught the workflow to write around the checker.
- Markup, accent macros and umlauts are normalized, so `\textbf{Snow}flake`
  cannot smuggle a forbidden skill past a substring test and
  `Universit\"at der Bundeswehr M\"unchen` is recognisably a school.
- Honest denials pass in English and German.
- The triple-list cadence hint knew only English "and", so it silently did
  nothing on every German package. Advisory only, and now labelled as such.

Verification
- The six counter-proofs from the report are caught: 9 errors on one synthetic
  document, where the old scanner found none.
- All six packages built under current rules scan clean — SBB, NVIDIA, RUAG,
  Kdo Cy, Citadel, Aker BP. Four false positives of my own were found and fixed
  on the way there; each is now a regression test.
- 13 older packages do report findings. Those look like genuine historical
  violations and are deliberately NOT fixed here: F17 governs archives and they
  must not be silently rewritten.
- `tests/test_document_scanner.py` added. Full suite 31/31 pass.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Qej5SBvhcbGXHwLqNWiskM

job_scout/state/application_cohort.json

@@ -223,13 +223,17 @@
         {
           "type": "rejected_no_interview",
           "date": "2026-08-27"
+        },
+        {
+          "type": "record_correction",
+          "date": "2026-09-11",
+          "details": "Application date CONFIRMED BY THE USER as 2026-07-29, settling the one-day conflict between the cohort record (07-29) and the CLAUDE.md session cell (07-30). The legacy import event above still quotes 07-30; it is preserved verbatim as the historical text and is not edited."
         }
       ],
       "cohort_id": "evidence-first-2026-01",
-      "unresolved": [
-        "Application date conflicts: cohort says 2026-07-29; CLAUDE says 2026-07-30. Preserve cohort date pending source confirmation."
-      ],
-      "fit_override_reason": "Legacy Adjacent classification retained; numeric score alone would suggest Core. Original rubric/override needs confirmation."
+      "fit_override_reason": "Recorded Adjacent although Evidence Fit 79 with a passing hard gate would score Core under the numeric rubric. The original rationale is NOT RECOVERABLE: the user was asked on 2026-09-11 and does not remember, and judged it immaterial because the role is closed and rejected. Deliberately NOT reclassified to Core - that would be the silent reclassification the review objected to, and it would rewrite a historical judgement to make the arithmetic tidy. The contemporaneous note points at a real reason rather than an error: the title-defining responsibility (authoring an enterprise framework) was only Adjacent, and application_strategy.md says a defining responsibility that is only Adjacent pushes a role down a class. Treated as a judgement recorded under the pre-2026-09-05 rubric, not as a data error. No future record may cite this one as precedent without stating its own reason.",
+      "rubric_version": "pre-2026-09-05-cohort",
+      "classification_note_for_stats": "Fit-class counts that include this row mix one pre-2026-09-05 judgement with post-rubric entries. Immaterial for a closed role; state it if the distribution is ever used as evidence."
     },
     {
       "company": "Google",