fix(config): replace TOKEN_SIGNING_SECRET entropy heuristic

[babs]

Summary

The previous distinct-byte heuristic on TOKEN_SIGNING_SECRET was structurally broken in both directions, surfaced by an external review pass:

• False positive. A real openssl rand -hex 32 value is drawn from a 16-symbol alphabet; expected number of distinct chars in 64 draws is ~15.75, so a genuinely random secret intermittently fell below the < 16 threshold and was rejected at startup under PROD_MODE=true. The README itself recommended this generator.
• False negative. Literals like `0123456789abcdef0123456789abcdef` have exactly 16 distinct chars and passed the gate despite trivially repeating with period 16. A patterned secret with no actual entropy past the first 16 bytes would clear the check.

Replacement: three-part shape check

1. All-same byte (period 1) → `is N copies of a single byte`.
2. Repeating period < len(b) → `repeats with period P`. Catches both the period-16 false-pass above and shorter shapes like `abcabc…`.
3. Tiny alphabet (< 8 distinct values) → catches the `aaaa…b`-shaped near-miss and uneven-run-length shapes (`aaaaabbbbbcccccddddd…`) that defeat both the period and all-same checks.

Threshold = 8 is the most defensive floor with zero practical false-positive rate on real random output: 32 hex chars over a 16-symbol alphabet expect ~14 distinct; base64 / raw expect more.

Behavioral envelope unchanged

Same two enforcement sites as before — warn path during secret parsing, hard-fail under `PROD_MODE`. Dev / single-replica deployments that intentionally use a weak secret keep working under `PROD_MODE=false` with a startup warning.

Tests

Test	What it pins
`TestLoad_SecretWeaknessWarning_PeriodicRepeat`	Headline regression for the famous false-pass: rejects `0123456789abcdef0123456789abcdef`.
`TestLoad_SecretWeaknessWarning_TinyAlphabet`	4-case boundary table: 2-distinct rejects, 7-distinct rejects (just below floor), 8-distinct passes (at floor, true 8-symbol non-periodic fixture), 15-distinct passes.
`TestLoad_SecretWeaknessWarning_NonPeriodic`	False-positive direction: frozen `openssl rand -hex 32`, base64-shaped, and alphanumeric-32 samples all pass.
`TestWeakSecretReason_FreshRandomNeverFlags`	9000-sample experiment per CI run (1000 iters × 3 encodings × 3 raw-byte lengths). Demonstrates zero false-positives on fresh `crypto/rand` output across raw, hex, and base64 shapes at lengths 32 / 48 / 64. A flake here is a real bug, not test noise.
`TestLoad_ProdMode_BlocksWeakSecret`	Now table-driven (3 shapes × 2 modes = 6 sub-tests).

The `setAllRequired` test fixture's `TOKEN_SIGNING_SECRET` moved from the literal famous false-pass (`0123456789abcdef…`) to a non-periodic real-random hex sample so every existing test runs against a secret the new gate accepts.

`distinctByteCount` removed entirely — no dead code, no temptation for a future maintainer to resurrect the broken heuristic.

Docs

`specs.md`, `README.md`, and `docs/configuration.md` all describe the new semantics and point at `manifests/scripts/generate-signing-secret.sh` as the canonical generator path.

Test plan

• CI green (`test`, `lint`, `keycloak-e2e`, `fuzz-smoke`, `manifest-prod`, `govulncheck`, `build`).
• `go test ./config/ -run "Secret|FreshRandom|TinyAlphabet" -v` passes (8000+ assertions land in the new tests).
• Manual sanity: a real `openssl rand -hex 32` value loads cleanly under `PROD_MODE=true` (was previously intermittent).