chore(release): v0.4.0, since LINKS-OPEN-REFS is a new check

Same precedent as v0.2.0 and v0.3.0: a new check code is a minor bump,
not a patch. Bundles the register row for repo #20, the schemaless-host
fix, the reference dedup and the new OK/SKIP outcomes on checkLinks.

Also records the register-freshness decision in CLAUDE.md: --refresh
runs HERE before every release, and consumers of the register are told
not to check freshness themselves. Twice running a newly published repo
was missing when a census ran, and the cost landed in a DIFFERENT repo
than the stale one — portfolio-optimiser's status line stayed put
through a round in which it fixed everything really wrong with it.
Two owners of one invariant is the drift this register exists to remove.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01NxP9N3p1fG6UYSB8rATBL6

README.md

@@ -10,7 +10,7 @@ checks that surface in one repository and reports what it finds.
 
 *AI-generated: all code produced by Claude Code through dialog-driven development.*
 
-![Version](https://img.shields.io/badge/version-0.3.0-blue)
+![Version](https://img.shields.io/badge/version-0.4.0-blue)
 ![Platform](https://img.shields.io/badge/platform-Claude_Code_Plugin-purple)
 ![Skills](https://img.shields.io/badge/skills-1-orange)
 ![License](https://img.shields.io/badge/license-MIT-lightgrey)
@@ -167,7 +167,7 @@ without that, a raw scan turns three dead names into about twenty.
 npm test
 ```
 
-116 tests over the pure classifiers. The reference fixtures are measured false
+129 tests over the pure classifiers. The reference fixtures are measured false
 positives, each with its expected verdict — the six that produced the
 three-outcome reference rule, plus the noise sources found by running the gate
 against a real repository: regexes inside code spans that are markdown links to