feat(diag): dogfood log-shipping to logs.sulkta.com

Ship scrubbed device logs to our self-hosted ingest (logs.sulkta.com) so a
'green in every offline gate, broken on-device' regression (the visionOS
playback outage) can be caught end-to-end. LogShipper POSTs a scrubbed log
ring + crash stacktrace on (a) an uncaught exception (chained to the previous
handler) and (b) a 'Send logs to Kayos' Settings button. Everything runs
through LogDump.scrubLine first (URLs/video-ids/emails/tokens redacted); the
server re-scrubs. device_id = a random per-install UUID (no hardware id).
Compile-DISABLED unless the STRAW_LOGS_INGEST_TOKEN build secret is set, so
contributor forks never phone home. build.yml passes the secret to the debug
build.

.forgejo/workflows/build.yml

@@ -88,6 +88,10 @@ jobs:
           STRAW_KEYSTORE_PASS: android
           STRAW_KEY_ALIAS: androiddebugkey
           STRAW_KEY_PASS: android
+          # Dogfood log-shipping: the app is compile-DISABLED unless this is set
+          # (empty ⇒ no crash handler, no network, Settings row hidden). Repo
+          # secret; absent on contributor forks so their builds don't phone home.
+          STRAW_LOGS_INGEST_TOKEN: ${{ secrets.STRAW_LOGS_INGEST_TOKEN }}
           # Keep the 4-ABI cross-compile off the container rootfs.
           CARGO_TARGET_DIR: ${{ github.workspace }}/cargo-target
         # ionice idle class + low nice so the build yields disk/CPU to the

strawApp/build.gradle.kts

@@ -10,6 +10,7 @@
  */
 
 import com.android.build.api.dsl.ApplicationExtension
+import java.util.Properties
 
 plugins {
     alias(libs.plugins.android.application)
@@ -36,6 +37,30 @@ configure<ApplicationExtension> {
         versionCode = STRAW_VERSION_CODE
         versionName = STRAW_VERSION_NAME
         resValue("string", "app_name", "Straw")
+
+        // Dogfood log-shipping token (feature/diag/LogShipper.kt). This repo
+        // is PUBLIC — the token must NEVER be committed, so it's injected at
+        // build time and defaults to "":
+        //   1. env var STRAW_LOGS_INGEST_TOKEN — CI: add a Forgejo Actions
+        //      repo secret STRAW_LOGS_INGEST_TOKEN and pass it in the
+        //      "Assemble debug APK" step's env block in
+        //      .forgejo/workflows/build.yml.
+        //   2. `straw.logs.ingest.token=…` in local.properties (gitignored)
+        //      for local dogfood builds.
+        // Empty token ⇒ LogShipper is fully disabled (no-op, zero network,
+        // no Settings row) — contributor builds neither break nor phone home.
+        val strawLogsToken: String = System.getenv("STRAW_LOGS_INGEST_TOKEN")
+            ?: run {
+                val props = Properties()
+                val f = rootProject.file("local.properties")
+                if (f.exists()) f.inputStream().use { props.load(it) }
+                props.getProperty("straw.logs.ingest.token")
+            }
+            ?: ""
+        // Escape so a pathological token can't break out of the generated
+        // String literal in BuildConfig.java.
+        val escapedToken = strawLogsToken.replace("\\", "\\\\").replace("\"", "\\\"")
+        buildConfigField("String", "STRAW_LOGS_TOKEN", "\"$escapedToken\"")
     }
 
     // Explicit signing so CI / release builds reuse ONE keystore instead of

strawApp/src/main/kotlin/com/sulkta/straw/StrawApp.kt

@@ -21,6 +21,7 @@ import com.sulkta.straw.data.SearchCache
 import com.sulkta.straw.data.Settings
 import com.sulkta.straw.data.Subscriptions
 import com.sulkta.straw.feature.dataimport.SettingsImport
+import com.sulkta.straw.feature.diag.LogShipper
 import com.sulkta.straw.feature.feed.FeedRefreshScheduler
 import com.sulkta.straw.feature.update.UpdateScheduler
 import com.sulkta.straw.feature.update.runUpdateCheck
@@ -90,6 +91,12 @@ class StrawApp : Application(), SingletonImageLoader.Factory {
 
     override fun onCreate() {
         super.onCreate()
+        // Dogfood crash shipping — wrap the default uncaught-exception
+        // handler FIRST so even an init crash below gets shipped. The
+        // wrapper always chains to the previous handler (normal crash
+        // behavior is untouched) and is a no-op in builds without an
+        // injected STRAW_LOGS_TOKEN.
+        LogShipper.installCrashHandler(this)
         // Path C-7: route Rust `log::*` calls into Android logcat under tag
         // "strawcore". Without this, every log line emitted from rustypipe /
         // strawcore is silently dropped, making playback regressions invisible

strawApp/src/main/kotlin/com/sulkta/straw/feature/diag/LogShipper.kt

@@ -0,0 +1,230 @@
+/*
+ * SPDX-FileCopyrightText: 2026 Sulkta
+ * SPDX-License-Identifier: GPL-3.0-or-later
+ *
+ * Dogfood log shipping — POST a scrubbed logcat window to Kayos's
+ * ingest service (logs.sulkta.com). Two triggers:
+ *
+ *   * crash  — a default-UncaughtExceptionHandler wrapper captures the
+ *              log ring + the throwable's stacktrace, attempts one
+ *              hard-capped best-effort send, then chains to the
+ *              previously-installed handler so the process still dies
+ *              (and Android still shows its crash dialog) normally.
+ *   * manual — the Settings → Diagnostics "Send logs to Kayos" button.
+ *
+ * PRIVACY: every line — including the crash stacktrace — goes through
+ * LogDump.scrubLine before leaving the device; the server re-scrubs,
+ * but the phone is the primary gate. device_id is a per-install random
+ * UUID (no ANDROID_ID / IMEI / hardware fingerprint). And the whole
+ * feature only exists in builds where BuildConfig.STRAW_LOGS_TOKEN was
+ * injected (see strawApp/build.gradle.kts) — for everyone else
+ * `enabled` is a compile-time false, so R8 strips the network path and
+ * a contributor build cannot phone home.
+ *
+ * DELIVERY: strictly best-effort. Any failure (offline, 401/413/429,
+ * timeout) is swallowed; nothing here may ever crash or block the app.
+ */
+
+package com.sulkta.straw.feature.diag
+
+import android.content.Context
+import com.sulkta.straw.BuildConfig
+import com.sulkta.straw.util.LogDump
+import com.sulkta.straw.util.runCatchingCancellable
+import java.io.PrintWriter
+import java.io.StringWriter
+import java.text.SimpleDateFormat
+import java.util.Date
+import java.util.Locale
+import java.util.TimeZone
+import java.util.UUID
+import java.util.concurrent.TimeUnit
+import kotlinx.coroutines.Dispatchers
+import kotlinx.coroutines.withContext
+import kotlinx.serialization.SerialName
+import kotlinx.serialization.Serializable
+import kotlinx.serialization.encodeToString
+import kotlinx.serialization.json.Json
+import okhttp3.MediaType.Companion.toMediaType
+import okhttp3.OkHttpClient
+import okhttp3.Request
+import okhttp3.RequestBody.Companion.toRequestBody
+
+object LogShipper {
+
+    /** Not a secret — only the bearer token is. */
+    private const val INGEST_URL = "https://logs.sulkta.com/ingest"
+
+    /** Server caps: 10k lines / 1 MB body. Stay well under both. */
+    private const val MAX_LINES = 2_000
+    private const val MAX_BODY_BYTES = 700_000
+
+    /** Hard cap on how long the crash path may delay process death. */
+    private const val CRASH_JOIN_MS = 4_000L
+
+    private const val PREFS = "straw_diag"
+    private const val KEY_DEVICE_ID = "device_id"
+
+    /**
+     * Compile-time gate: builds without an injected token have no
+     * shipping at all — no network, no crash handler, no Settings row.
+     */
+    val enabled: Boolean
+        get() = BuildConfig.STRAW_LOGS_TOKEN.isNotEmpty()
+
+    // Short timeouts — this is telemetry, not payload. callTimeout is
+    // the absolute ceiling on the whole request so nothing here can
+    // hold a WorkManager-style wakelock pattern open.
+    private val http: OkHttpClient by lazy {
+        OkHttpClient.Builder()
+            .connectTimeout(5, TimeUnit.SECONDS)
+            .writeTimeout(5, TimeUnit.SECONDS)
+            .readTimeout(5, TimeUnit.SECONDS)
+            .callTimeout(10, TimeUnit.SECONDS)
+            .build()
+    }
+
+    @Serializable
+    private data class Envelope(
+        @SerialName("device_id") val deviceId: String,
+        @SerialName("app_version") val appVersion: String,
+        @SerialName("sent_at") val sentAt: String,
+        val reason: String,
+        val lines: List<String>,
+    )
+
+    /**
+     * Stable per-install id: `straw-<random uuid>`, minted once and
+     * persisted. Deliberately NOT ANDROID_ID/IMEI — it identifies an
+     * install for log correlation, never the hardware or the person.
+     * commit() (not apply()) because both call paths are already on a
+     * background thread and the crash path dies right after — an
+     * async write would lose the id and mint a new one per crash.
+     */
+    private fun deviceId(context: Context): String {
+        val sp = context.getSharedPreferences(PREFS, Context.MODE_PRIVATE)
+        sp.getString(KEY_DEVICE_ID, null)?.let { return it }
+        val id = "straw-${UUID.randomUUID()}"
+        sp.edit().putString(KEY_DEVICE_ID, id).commit()
+        return id
+    }
+
+    private fun isoNowUtc(): String =
+        SimpleDateFormat("yyyy-MM-dd'T'HH:mm:ss.SSS'Z'", Locale.US)
+            .apply { timeZone = TimeZone.getTimeZone("UTC") }
+            .format(Date())
+
+    /**
+     * Manual trigger (Settings). Captures + scrubs the log ring and
+     * POSTs it. Returns true iff the server 2xx'd — the UI shows a
+     * toast either way; failures never throw.
+     */
+    suspend fun send(context: Context, reason: String): Boolean =
+        withContext(Dispatchers.IO) {
+            if (!enabled) return@withContext false
+            runCatchingCancellable {
+                postLines(
+                    context = context,
+                    reason = reason,
+                    lines = LogDump.captureScrubbedLinesBlocking(MAX_LINES),
+                )
+            }.getOrDefault(false)
+        }
+
+    /**
+     * Wrap the process-default uncaught-exception handler. Ship-then-
+     * chain: whatever happens in our path (including a throw), the
+     * previous handler ALWAYS runs, so the app still crashes/reports
+     * exactly as before. No-op without a token.
+     */
+    fun installCrashHandler(context: Context) {
+        if (!enabled) return
+        val appContext = context.applicationContext
+        val previous = Thread.getDefaultUncaughtExceptionHandler()
+        Thread.setDefaultUncaughtExceptionHandler { thread, throwable ->
+            try {
+                shipCrash(appContext, throwable)
+            } catch (_: Throwable) {
+                // Never let diagnostics mask the real crash.
+            }
+            previous?.uncaughtException(thread, throwable)
+        }
+    }
+
+    @Volatile
+    private var crashShipStarted = false
+
+    /**
+     * Crash-path send. The crashing thread may be main (network there
+     * throws NetworkOnMainThreadException) and is about to die either
+     * way, so the capture+POST runs on a dedicated worker thread and
+     * we join with a hard cap — the death is delayed at most
+     * [CRASH_JOIN_MS], never hung.
+     */
+    private fun shipCrash(context: Context, throwable: Throwable) {
+        // One attempt per process — also breaks recursion if the ship
+        // path itself crashes on another thread mid-flight.
+        if (crashShipStarted) return
+        crashShipStarted = true
+        val worker = Thread {
+            try {
+                val trace = StringWriter().also {
+                    throwable.printStackTrace(PrintWriter(it))
+                }.toString()
+                    .split('\n')
+                    .take(400)
+                    .map(LogDump::scrubLine)
+                val ring = LogDump.captureScrubbedLinesBlocking(MAX_LINES)
+                postLines(
+                    context = context,
+                    reason = "crash",
+                    lines = ring + "--- uncaught exception ---" + trace,
+                )
+            } catch (_: Throwable) {
+                // Best-effort only.
+            }
+        }
+        worker.name = "straw-crash-ship"
+        worker.isDaemon = true
+        worker.start()
+        try {
+            worker.join(CRASH_JOIN_MS)
+        } catch (_: InterruptedException) {
+            Thread.currentThread().interrupt()
+        }
+    }
+
+    /**
+     * Build the envelope and POST. `lines` MUST already be scrubbed
+     * (both call paths run everything through LogDump.scrubLine).
+     * Trims oldest-first to stay under the server's 1 MB cap — one
+     * bounded request, no chunking needed at our line cap.
+     */
+    private fun postLines(context: Context, reason: String, lines: List<String>): Boolean {
+        var budget = MAX_BODY_BYTES
+        val kept = ArrayList<String>(lines.size)
+        for (i in lines.indices.reversed()) {
+            // +8 ≈ JSON quotes/comma + escaping slack per line.
+            val cost = lines[i].toByteArray(Charsets.UTF_8).size + 8
+            if (cost > budget) break
+            budget -= cost
+            kept.add(lines[i])
+        }
+        kept.reverse()
+        val envelope = Envelope(
+            deviceId = deviceId(context),
+            appVersion = "${BuildConfig.VERSION_NAME} (${BuildConfig.VERSION_CODE})",
+            sentAt = isoNowUtc(),
+            reason = reason,
+            lines = kept,
+        )
+        val body = Json.encodeToString(envelope)
+            .toRequestBody("application/json; charset=utf-8".toMediaType())
+        val request = Request.Builder()
+            .url(INGEST_URL)
+            .header("Authorization", "Bearer ${BuildConfig.STRAW_LOGS_TOKEN}")
+            .post(body)
+            .build()
+        return http.newCall(request).execute().use { it.isSuccessful }
+    }
+}

strawApp/src/main/kotlin/com/sulkta/straw/feature/settings/SettingsScreen.kt

@@ -66,6 +66,7 @@ import com.sulkta.straw.data.Settings
 import com.sulkta.straw.data.ThemeMode
 import com.sulkta.straw.feature.dataimport.ImportResult
 import com.sulkta.straw.feature.dataimport.SettingsImport
+import com.sulkta.straw.feature.diag.LogShipper
 import com.sulkta.straw.feature.feed.SubscriptionFeedViewModel
 import com.sulkta.straw.feature.search.SearchViewModel
 import com.sulkta.straw.util.LogDump
@@ -774,6 +775,39 @@ fun SettingsScreen() {
             Text(if (logDumping) "Exporting…" else "Export logs…")
         }
 
+        // Dogfood log shipping — only rendered in builds with an
+        // injected ingest token (LogShipper.enabled is compile-time
+        // false otherwise, so contributor builds show no phantom row).
+        if (LogShipper.enabled) {
+            Spacer(modifier = Modifier.height(16.dp))
+            Text(
+                "Or send the same scrubbed log dump straight to Kayos " +
+                    "(logs.sulkta.com) — no share sheet. URLs, video ids, " +
+                    "and anything token-shaped are redacted on-device first.",
+                style = MaterialTheme.typography.bodySmall,
+                color = MaterialTheme.colorScheme.onSurfaceVariant,
+            )
+            Spacer(modifier = Modifier.height(12.dp))
+            var logSending by remember { mutableStateOf(false) }
+            OutlinedButton(
+                enabled = !logSending,
+                onClick = {
+                    logSending = true
+                    scope.launch {
+                        val ok = LogShipper.send(context, reason = "manual")
+                        logSending = false
+                        Toast.makeText(
+                            context,
+                            if (ok) "Logs sent — thanks!" else "Couldn't reach the log server",
+                            Toast.LENGTH_LONG,
+                        ).show()
+                    }
+                },
+            ) {
+                Text(if (logSending) "Sending…" else "Send logs to Kayos")
+            }
+        }
+
         Spacer(modifier = Modifier.height(32.dp))
         Text(
             "Import from NewPipe / Tubular",

strawApp/src/main/kotlin/com/sulkta/straw/util/LogDump.kt

@@ -96,10 +96,46 @@ object LogDump {
     }
 
     /**
-     * Pre-redact known credential-shaped substrings before they hit
-     * disk. Cheap line-level pass — adversarial-perfect would need a
-     * URL parser, but the regex approach catches every documented
-     * leak vector at zero allocation cost.
+     * Pull recent logcat, scrub every line, keep only the newest
+     * [maxLines]. BLOCKING — callers must already be off the main
+     * thread (LogShipper runs on Dispatchers.IO; the crash handler
+     * runs this on its own short-lived worker thread, where a suspend
+     * fun can't be awaited without an event loop).
+     *
+     * Best-effort by design: a logcat exec failure returns whatever
+     * was read (possibly empty) instead of throwing, so the crash
+     * shipper can still deliver the stacktrace.
+     */
+    fun captureScrubbedLinesBlocking(maxLines: Int): List<String> {
+        val ring = ArrayDeque<String>()
+        runCatching {
+            val pid = Process.myPid()
+            val cmd = arrayOf("logcat", "-d", "-v", "threadtime", "--pid=$pid")
+            val proc = ProcessBuilder(*cmd).redirectErrorStream(true).start()
+            proc.inputStream.bufferedReader().useLines { lines ->
+                lines.forEach { line ->
+                    if (ring.size == maxLines) ring.removeFirst()
+                    ring.addLast(line)
+                }
+            }
+            proc.waitFor()
+        }
+        // Scrub only the retained window — cheaper than scrubbing lines
+        // that get dropped, and nothing raw ever leaves this function.
+        return ring.map(::scrubLine)
+    }
+
+    /**
+     * Pre-redact known credential- and identity-shaped substrings
+     * before they hit disk or the wire. Cheap line-level pass —
+     * adversarial-perfect would need a URL parser, but the regex
+     * approach catches every documented leak vector at zero
+     * allocation cost.
+     *
+     * BIAS: over-redaction. These lines leave the device (share sheet
+     * + LogShipper → logs.sulkta.com), so a false positive costs a
+     * little debug signal while a false negative leaks what someone
+     * watched. Anything URL-, token-, id-, or address-shaped goes.
      *
      * Public so error-handler call sites (PlayerScreen / VideoDetail
      * `playbackError`) can scrub Media3's `PlaybackException.message`
@@ -109,28 +145,112 @@ object LogDump {
      */
     fun scrubLine(line: String): String {
         var s = line
-        // Pre-signed googlevideo URLs: keep host visible, drop path+query.
+        // Pre-signed googlevideo URLs: keep the host label visible, drop
+        // host prefix + path + query (session-bound streaming creds).
         s = GOOGLEVIDEO_URL_RE.replace(s, "https://<host>.googlevideo.com/<scrubbed>")
+        // Credential-shaped headers / key-value pairs, wherever they
+        // appear: `Authorization: Bearer x`, `cookie: …`, `api_key=…`.
+        s = BEARER_RE.replace(s, "$1 <scrubbed>")
+        s = CRED_KV_RE.replace(s, "$1: <scrubbed>")
         // Long, distinctive token names — match anywhere.
         s = SIGNED_PARAM_LONG_RE.replace(s, "$1=<scrubbed>")
         // Short single-letter / two-letter tokens — require `[?&]`
         // immediately before to avoid eating innocent counters.
         s = SIGNED_PARAM_SHORT_RE.replace(s, "$1$2=<scrubbed>")
+        // ANY remaining URL: keep scheme+host (routing/debug signal,
+        // not PII), drop path + query — that's where watch?v=…,
+        // /vi/<id>/ thumbnails, search terms, etc. live.
+        s = URL_TAIL_RE.replace(s, "$1/<scrubbed>")
+        // Schemeless YT links pasted into log messages ("youtu.be/xyz").
+        s = SCHEMELESS_YT_RE.replace(s, "$1/<scrubbed>")
+        // Emails.
+        s = EMAIL_RE.replace(s, "<email>")
+        // YouTube channel + playlist ids (watch behavior).
+        s = CHANNEL_ID_RE.replace(s, "<channelId>")
+        s = PLAYLIST_ID_RE.replace(s, "<playlistId>")
+        // Bare 11-char video-id-shaped tokens ("dQw4w9WgXcQ") logged
+        // outside any URL (rustypipe/strawcore do this). Requiring a
+        // digit/_/- inside the run spares ordinary 11-letter words;
+        // real ids virtually always contain one. Pure-alpha ids are
+        // still caught earlier when keyed (v=…) or inside a URL.
+        s = VIDEO_ID_CANDIDATE_RE.replace(s) { m ->
+            if (m.value.any { it.isDigit() || it == '_' || it == '-' }) "<videoId>" else m.value
+        }
+        // Long high-entropy runs (hashes, visitor data, unlabeled
+        // tokens): 20+ [A-Za-z0-9_-] chars containing a digit.
+        s = LONG_TOKEN_RE.replace(s, "<token>")
+        // IP addresses (v4 + v6 — v6 patterns are shaped so threadtime
+        // HH:MM:SS timestamps can never match).
+        s = IPV4_RE.replace(s, "<ip>")
+        s = IPV6_FULL_RE.replace(s, "<ip>")
+        s = IPV6_COMPRESSED_RE.replace(s, "<ip>")
         return s
     }
 
     private val GOOGLEVIDEO_URL_RE = Regex(
         """https?://[a-zA-Z0-9.-]*googlevideo\.com/\S+""",
     )
+    // `Authorization: Bearer <jwt>` and friends. The token charset is
+    // the RFC 6750 b64token alphabet.
+    private val BEARER_RE = Regex(
+        """\b(bearer)\s+[A-Za-z0-9._~+/=-]{4,}""",
+        RegexOption.IGNORE_CASE,
+    )
+    // Credential-y names followed by `: value` or `= value` (optional
+    // closing quote for JSON `"token":"…"` shapes). Values may be
+    // quoted strings or a bare token.
+    private val CRED_KV_RE = Regex(
+        """\b(authorization|proxy-authorization|cookie|set-cookie|x-goog-[a-z\-]+|x-youtube-[a-z\-]+|api[-_]?key|access[-_]?token|refresh[-_]?token|id[-_]?token|client[-_]?secret|secret|token|auth|session[-_]?id|passw(?:or)?d|pwd|mnemonic|private[-_]?key)"?\s*[:=]\s*("[^"]*"|\S+)""",
+        RegexOption.IGNORE_CASE,
+    )
     // Long tokens are unique enough to match anywhere. Short tokens
-    // (n, mn, ms, mo, pl, ip, ei) require `[?&]` immediately before
-    // so we don't redact innocuous `n=42` counters from other libs.
+    // (n, mn, ms, … v, id, q) require `[?&]` immediately before so we
+    // don't redact innocuous `n=42` counters from other libs. `v`,
+    // `id`, `list`, `q`, `t` are the YouTube identity/search params.
     private val SIGNED_PARAM_LONG_RE = Regex(
-        """\b(signature|sparams|lsig|cpn|expire|pot|sig|key)=([^&\s"']+)""",
+        """\b(signature|sparams|lsig|cpn|expire|pot|sig|key|videoId|video_id|docid|search_query|query)=([^&\s"']+)""",
         RegexOption.IGNORE_CASE,
     )
     private val SIGNED_PARAM_SHORT_RE = Regex(
-        """([?&])(n|mn|ms|mo|pl|ip|ei)=([^&\s"']+)""",
+        """([?&])(n|mn|ms|mo|pl|ip|ei|v|id|list|q|t)=([^&\s"']+)""",
         RegexOption.IGNORE_CASE,
     )
+    // Any http(s) URL: capture scheme + host(+port), scrub the rest.
+    // Host charset can't contain `<`, so the googlevideo replacement
+    // above is never re-matched.
+    private val URL_TAIL_RE = Regex(
+        """(https?://[A-Za-z0-9.-]+(?::\d+)?)[/?#][^\s"'<>]*""",
+    )
+    // youtu.be/… + youtube.com/… without a scheme. The lookbehind
+    // skips hosts already handled as part of a full URL (preceded by
+    // `/` or a subdomain dot).
+    private val SCHEMELESS_YT_RE = Regex(
+        """(?<![/@.\w])((?:www\.|m\.|music\.)?youtube\.com|youtu\.be)[/?][^\s"'<>]*""",
+    )
+    private val EMAIL_RE = Regex(
+        """[A-Za-z0-9._%+-]+@[A-Za-z0-9.-]+\.[A-Za-z]{2,}""",
+    )
+    private val CHANNEL_ID_RE = Regex(
+        """(?<![A-Za-z0-9_-])UC[A-Za-z0-9_-]{22}(?![A-Za-z0-9_-])""",
+    )
+    private val PLAYLIST_ID_RE = Regex(
+        """(?<![A-Za-z0-9_-])(?:PL|UU|LL|RD|OLAK5uy_)[A-Za-z0-9_-]{10,}(?![A-Za-z0-9_-])""",
+    )
+    private val VIDEO_ID_CANDIDATE_RE = Regex(
+        """(?<![A-Za-z0-9_-])[A-Za-z0-9_-]{11}(?![A-Za-z0-9_-])""",
+    )
+    private val LONG_TOKEN_RE = Regex(
+        """(?<![A-Za-z0-9_-])(?=[A-Za-z_-]*\d)[A-Za-z0-9_-]{20,}(?![A-Za-z0-9_-])""",
+    )
+    private val IPV4_RE = Regex(
+        """(?<!\d)(?:\d{1,3}\.){3}\d{1,3}(?!\d)""",
+    )
+    // Full-form v6 needs 6-8 hex groups — `HH:MM:SS` (3 groups) can't
+    // match. Compressed form requires a literal `::`.
+    private val IPV6_FULL_RE = Regex(
+        """(?<![0-9A-Fa-f:])(?:[0-9A-Fa-f]{1,4}:){5,7}[0-9A-Fa-f]{1,4}(?![0-9A-Fa-f:])""",
+    )
+    private val IPV6_COMPRESSED_RE = Regex(
+        """(?<![0-9A-Fa-f:])(?:[0-9A-Fa-f]{1,4}:){1,6}:(?:[0-9A-Fa-f]{1,4}(?::[0-9A-Fa-f]{1,4}){0,5})?(?![0-9A-Fa-f:])""",
+    )
 }