Audit review (rounds 1-31): fix confirmed issues, remove dead code, harden tests

Security/correctness fixes:
- fail-open token validation preserved with lenient-audience; access token that
  fails verification and has no ID token to corroborate now fails closed
- token_manager blacklist-expiry: fix impossible tautology (x>24h && x<24h)
- constant-time OAuth state compare; clear state/nonce/code_verifier on
  callback token-exchange failure to prevent replay of half-failed flows
- StopAllTasks gated on last-instance close
- background-task panic recovery; remove task-registry counter leak
- circuit breaker half-open: release probe slot; CB + per-session outcome
  recorded once per refresh operation, not once per waiter
- Redis connection pool: handle nil receive when pool closed concurrently
- FileStore: atomic temp-write + rename
- errors_recovery health-check data race -> atomic

Test-isolation / hardening:
- reset helpers for Once-pinned global cache manager and transport pool
- coverage: utils 70->93% (redirect URL security contract); fuzz targets for
  bearer parsers; concurrent ServeHTTP race tests for the shared request path

Dead-code removal (verified orphaned, build+suite green): internal/providers
(30 files, 0 importers), internal/compat, internal/features, input_validation.

Verified: govulncheck, go mod tidy, yaegi-validate, cross-compile (darwin+
linux), golangci-lint (new-from-rev clean), 3x full race+shuffle green.

internal/recovery/circuit_breaker.go

@@ -177,8 +177,18 @@ func (cb *CircuitBreaker) recordSuccess() {
 	state := CircuitBreakerState(atomic.LoadInt32(&cb.state))
 
 	// #nosec G115 -- SuccessThreshold is a small config value that fits in int32
-	if state == CircuitBreakerHalfOpen && successes >= int32(cb.config.SuccessThreshold) {
-		cb.transitionToClosed()
+	if state == CircuitBreakerHalfOpen {
+		if successes >= int32(cb.config.SuccessThreshold) {
+			cb.transitionToClosed() // resets halfOpenRequests to 0
+		} else {
+			// Release the probe slot so the half-open gate can accept the
+			// next request. Without this, halfOpenRequests stays pinned at
+			// MaxRequests once the first batch of probes completes, every
+			// later request is rejected, and consecutiveSuccesses can never
+			// reach SuccessThreshold if it exceeds MaxRequests — leaving the
+			// circuit stuck half-open forever.
+			atomic.AddInt32(&cb.halfOpenRequests, -1)
+		}
 	}
 }