Test-quality audit: fix 2 prod bugs, strengthen weak tests, add mutation/fuzz/CI tooling (#5345)

* test(audit): add gremlins/rapid/coverage tooling + AUDIT.md scaffold

* test(audit): hygiene sweep (race-clean except logger global; Finding #2) + smell inventory

* test(audit): cover untested error/edge branches (TLS proxy+pin, migration tag cleanup=Finding #1)

* test(audit): strengthen internal/sub link tests (dedup key, TLS/Reality mapping, clash well-formedness)

* test(audit): property (rapid) + fuzz tests for joinHostPort/userinfo/pin/ParseLink

* test(audit): tighten frontend subSortIndex rejection assertions + wire coverage

* ci(audit): add shuffle gate + non-blocking race job (Finding #2) + fuzz-smoke; document mutation policy

* chore(audit): gitignore frontend coverage output

* test(audit): exhaustive whole-repo pass — strengthen 5 weak/fake tests (netproxy, CSP, modal per-protocol loops, schema coercions)

* docs(contributing): add Testing section (conventions, race/shuffle, fuzz, mutation policy); drop AUDIT.md ledger

* fix(logger,migration): guard logBuffer with mutex; execute legacy tag cleanup (tx.Exec); make CI race gate blocking

* ci(mutation): add nightly scoped gremlins workflow (informational artifacts)

* test(audit): strengthen runtime tests — baseURL scheme/port bounds, isNonEmptySlice, trafficReset

* test(audit): strengthen clash tests — reality field mapping + tcp-header validation

* test(audit): runtime — egress-proxy + content-type tests; drop redundant bp=='' branch

* test(audit): strengthen link parser/helper tests (defaultPort, splitComma, base64, canonicalQuery, tls/reality/transport mapping)

* test(audit): strengthen sub/xray/common/netsafe/mtproto/config/middleware tests (kill surviving mutants)

* test(audit): raise timeout on protocol-iteration modal tests (heavy re-renders, slow on CI)

* fix(logger): GetLogs returns at most c entries (off-by-one fix; addresses PR review)

* perf(logger): snapshot logBuffer under lock so GetLogs doesn't block logging; clarify fuzz-seed docs (addresses PR review)

CONTRIBUTING.md

@@ -236,6 +236,49 @@ For deeper notes on the frontend toolchain see [`frontend/README.md`](frontend/R
 | `internal/config/` | Environment-variable helpers, paths, defaults |
 | `x-ui/` | **Runtime data** — db, logs, xray binary, geo files (gitignored) |
 
+## Testing
+
+Tests live next to the code (`foo.go` ↔ `foo_test.go`); frontend specs and golden fixtures live in `frontend/src/test/`.
+
+### Go conventions
+
+- **Stdlib `testing` only** — no testify. Table-driven with `t.Run` subtests and `t.Helper()` on helpers.
+- **Assert the contract, not internals.** Pin the exact value / typed error / emitted string — not `err != nil` or `len > 0`. A test that still passes when the behavior is broken is worse than no test.
+- **Real dependencies over mocks.** Get a throwaway DB with `database.InitDB(filepath.Join(t.TempDir(), "x-ui.db"))` + `t.Cleanup(func() { _ = database.CloseDB() })` (Windows-safe), and use `httptest` servers for HTTP. The `internal/sub` suite's `initSubDB(t)` is the template.
+
+### Running
+
+| Goal | Command |
+|------|---------|
+| Standard run | `go test ./...` |
+| Hygiene — data races + order-dependence | `go test -race -shuffle=on -count=1 ./...` (`-race` needs the C compiler from Prerequisites) |
+| Coverage gaps | `go test -coverprofile=cov.out ./<pkg>/... && go tool cover -func=cov.out` |
+| Fuzz a parser briefly | `go test -run '^$' -fuzz 'FuzzName$' -fuzztime=30s ./<pkg>/...` |
+
+Frontend: `cd frontend && npm run test` (vitest), or `npm run test -- --coverage`.
+
+### Property and fuzz tests
+
+Input-heavy or pure logic (link builders, parsers, decoders) is also covered by **property tests** (`pgregory.net/rapid`) and **native fuzz targets** (`go test -fuzz`). A fuzz target's **seed corpus** (its inline `f.Add` cases plus any `testdata/fuzz` entries) runs as ordinary subtests under a plain `go test` — no `-fuzz` flag needed — so CI's normal test job exercises the seeds; the time-boxed *fuzzing* exploration (`-fuzz=...`) runs separately as the `fuzz-smoke` job.
+
+### Mutation testing (optional, manual)
+
+[gremlins](https://github.com/go-gremlins/gremlins) checks whether tests actually fail when the code is mutated — a surviving (`LIVED`) mutant means a weak test. It is **slow**, so run it **scoped per package**, never repo-wide or per-commit:
+
+```bash
+go install github.com/go-gremlins/gremlins/cmd/gremlins@latest
+gremlins unleash ./internal/sub/
+gremlins unleash -E 'server\.go|xray\.go|inbound\.go|client_bulk\.go|inbound_traffic\.go|.*_postgres_test\.go' ./internal/web/service/
+```
+
+Treat each survivor as one of: a weak test (strengthen it), dead code (remove it), or an equivalent mutant (unkillable — leave it). Don't write a test purely to kill a mutant if it doesn't reflect real behavior.
+
+CI runs this for you nightly (and on demand) via `.github/workflows/mutation.yml` — scoped per package, results uploaded as artifacts. It is **informational**, not a gate (no thresholds), so check the reports when hardening a suite rather than waiting for a red build.
+
+### CI
+
+`.github/workflows/ci.yml` runs per PR: `go-test` (with `-shuffle -count=1`), a `race` job (`-race -shuffle -count=1`), a `fuzz-smoke` job on the critical parsers, and the frontend `typecheck`/`lint`/`test`/`build`. Snapshots are regression guards — regenerate them (`npx vitest run -u`) only for intentional output changes, never to make a red test green.
+
 ## Sending a pull request
 
 1. Branch off `main` (e.g. `feat/short-description`).