dsh-remote 0.8.72 → 0.8.74

package/lib/web-attach.js

@@ -1,850 +1,850 @@
-// dsh-remote — "remote DSH web UI" attach sessions (issue #46).
-//
-// The user's intent behind that issue is "use the local DSH as a CLIENT for a
-// DSH running on another machine": they do NOT want the remote to expose an
-// inbound listener. DSH makes that impossible anyway — `dsh-web-app` hard-
-// rejects `--host 0.0.0.0` (startup.js) because it would put remote code
-// execution on the network, and `dsh-host-webserver` only accepts the
-// 127.0.0.1/0.0.0.0 literals.
-//
-// So this module inverts the direction, exactly like the rest of the plugin:
-// we SSH OUT, start (or reuse) a loopback-bound `dsh web` on the remote, and
-// carry its socket back over our own SSH connection with a local TCP forward.
-// The browser then talks to 127.0.0.1:<localPort> on this machine.
-//
-// Verified end-to-end on a real Linux host (dsh 0.1.5-rc.2): a bare TCP forward
-// is sufficient — the SPA, its assets, AND the `/api/remote.mux` WebSocket
-// upgrade all transit it (HTTP/1.1 101), with no reverse proxy and no host
-// rewriting. That works because DSH binds its auth cookie to the request
-// authority, so a cookie minted against `127.0.0.1:<localPort>` is exactly what
-// the browser then sends back to that same authority.
-//
-// Three details drive the design and are worth stating once:
-//
-//  1. The launch token CANNOT be read from anywhere but stdout. It is a
-//     `randomBytes` value held in a process-local WeakMap and printed on one
-//     line (`dsh web: http://127.0.0.1:<port>/?token=…`). No file, no env var.
-//     Therefore the token is captured by redirecting the remote process's
-//     output to a log file and parsing that line out of it.
-//  2. `--port 0` lets the remote OS pick a free port, and the same printed line
-//     reports which port it chose — so we never have to guess or probe for a
-//     free port on the remote.
-//  3. The pairing that must survive is (token, port) from ONE launch. Reading
-//     them from one line is what keeps them consistent.
-//
-// Everything here is read-only with respect to the remote filesystem except
-// the log file it creates under the remote `$HOME`.
-import net from 'node:net'
-
-/** Remote log file per attach session; `$HOME` is expanded by the remote shell. */
-const LOG_NAME = '.dsh-remote-web'
-
-/**
- * Parse the `dsh web:` startup line out of a remote log.
- *
- * The line looks like:
- *   dsh web: http://127.0.0.1:43129/?token=Iy7FKuPSI5qJKJUFDyah4B6i8Ul6_hvspiJhiWufYt8
- * and may be followed by a LAN suffix when the remote bound 0.0.0.0 (it never
- * will here — we never pass `--host` — but tolerate it rather than break).
- *
- * @param {string} text - the remote log contents.
- * @returns {{url: string, port: number, token: string}|null} parsed facts, or
- *   null when the line has not appeared yet (still booting) or is malformed.
- */
-export function parseWebStart(text) {
-  if (typeof text !== 'string' || text === '') return null
-  // Take the LAST match: a reused log file may hold an earlier session's line,
-  // and the newest one is the one this launch produced.
-  const matches = [...text.matchAll(/dsh web:\s*(http:\/\/[^\s]+)/g)]
-  if (matches.length === 0) return null
-  const url = matches[matches.length - 1][1]
-  let parsed
-  try {
-    parsed = new URL(url)
-  } catch {
-    return null
-  }
-  const port = Number(parsed.port)
-  const token = parsed.searchParams.get('token') || ''
-  if (!Number.isInteger(port) || port < 1 || port > 65535) return null
-  if (!token) return null
-  return { url, port, token }
-}
-
-/** True when the log proves the remote process died instead of booting. */
-export function launchFailed(text) {
-  return typeof text === 'string' && text.includes(DEAD_MARKER)
-}
-
-/** Marker the remote launcher echoes when the process exits before printing. */
-const DEAD_MARKER = '__DSH_REMOTE_DIED__'
-/** Marker separating the parsed line from the raw log tail for diagnostics. */
-export const TAIL_MARKER = '__DSH_TAIL__'
-/** Marker carrying the remote PID so the session can stop its own process. */
-export const PID_MARKER = '__DSH_PID__='
-/** Random ASCII tag so two attach sessions never share a remote log file. */
-function randomTag() {
-  return Math.random().toString(36).slice(2, 8) + Date.now().toString(36).slice(-4)
-}
-
-/**
- * Build the one-shot remote command that launches `dsh web` detached and waits
- * for its startup line.
- *
- * `setsid` + `nohup` + redirected stdio are all required: the SSH channel closes
- * as soon as this shell exits, and without detaching the SIGHUP would kill the
- * server we just started. The loop is bounded so a wedged remote cannot hang the
- * call past the caller's own timeout, and it reports either the parsed line or
- * the log tail — never silence.
- *
- * @param {object} spec - launch parameters.
- * @param {boolean} [spec.omitNoOpen] - drop `--no-open`. Older DSH (measured on
- *   0.1.0-rc.6) has no such flag and exits with "unknown option '--no-open'";
- *   the caller retries with this set rather than assuming a flag exists.
- * @returns {string} a POSIX shell command line (no newlines, ASCII only).
- */
-export function buildLaunchCommand({ command, profile, logPath, remotePort = 0, dshHome = '', waitSeconds = 45, omitNoOpen = false }) {
-  const env = dshHome ? `DSH_HOME="${dshHome}" ` : ''
-  const noOpen = omitNoOpen ? '' : ' --no-open'
-  const args = `--profile ${profile} --port ${Number(remotePort) || 0}${noOpen}`
-  const limit = Number(waitSeconds) || 45
-  // The loop is ONE shell element: `do` is a keyword and `do;` is a syntax error
-  // in POSIX sh, so it must be glued to the first command of the body rather
-  // than separated like the other statements.
-  const loop = `while [ $i -lt ${limit} ]; do grep -q 'token=' "$LOG" 2>/dev/null && break`
-    + `; kill -0 $P 2>/dev/null || { echo ${DEAD_MARKER}; break; }`
-    + `; sleep 1; i=$((i+1)); done`
-  return [
-    `LOG="$HOME/${logPath}"`,
-    `: > "$LOG"`,
-    // `command -v setsid` guards hosts without util-linux (macOS); plain nohup
-    // still detaches from the channel's HUP there.
-    `command -v setsid >/dev/null 2>&1 && SET=setsid || SET=`,
-    `${env}$SET nohup ${command} ${args} >"$LOG" 2>&1 </dev/null & P=$!`,
-    `i=0`,
-    loop,
-    `echo "${PID_MARKER}$P"`,
-    `grep -m1 'dsh web:' "$LOG" 2>/dev/null`,
-    `echo ${TAIL_MARKER}`,
-    `tail -20 "$LOG" 2>/dev/null`,
-  ].join('; ')
-}
-
-/** Extract the remote PID from launcher output (0 when absent). */
-export function parseRemotePid(output) {
-  const m = /__DSH_PID__=(\d+)/.exec(String(output || ''))
-  return m ? Number(m[1]) : 0
-}
-
-/**
- * Recognize a known remote-side reason the web surface could not start, and
- * return an actionable sentence for it.
- *
- * The point is to keep an operator from reading "the remote DSH did not report a
- * startup token" and concluding the *plugin* is broken. Measured cause: dsh
- * ≤ 0.1.0-rc.6 depends on `node-pty@1.1.0`, whose published tarball ships
- * darwin/win32 prebuilds but **no `linux-x64`**, so `dsh web` cannot boot on
- * Linux at all — it dies loading the native module. 0.1.5-rc.2 moved to
- * `node-pty@1.2.0-beta.15`, which does ship Linux builds.
- *
- * @param {string} output - the remote log tail.
- * @returns {string} a hint, or '' when nothing is recognized.
- */
-export function explainLaunchFailure(output) {
-  const text = String(output || '')
-  if (/pty\.node|Failed to load native module/i.test(text)) {
-    return ' The remote could not load node-pty\'s native module. Known cause: dsh <= 0.1.0-rc.6'
-      + ' depends on node-pty@1.1.0, which publishes no linux-x64 prebuild, so `dsh web` cannot start'
-      + ' on Linux at all. Upgrade the remote dsh to 0.1.5-rc.2 or newer (node-pty@1.2.0-beta.15 ships'
-      + ' the Linux build), or point `webAttachCommand` at a newer dsh installed elsewhere.'
-  }
-  if (/is intentionally not supported|--host 0\.0\.0\.0/i.test(text)) {
-    return ' The remote dsh refused an all-interfaces bind. Attach deliberately never passes `--host`;'
-      + ' check `webAttachCommand` does not add one.'
-  }
-  if (/command not found|No such file or directory|not found/i.test(text)) {
-    return ' The remote shell could not run the dsh command. If the remote dsh is not on the SSH login'
-      + ' PATH (a version manager or private prefix), set `webAttachCommand` to its absolute path.'
-  }
-  return ''
-}
-
-/**
- * Build the command that stops the remote process THIS session started.
- *
- * Two things make this less obvious than it looks, both found by running it
- * against a real host:
- *   1. The log path cannot identify the process — it appears only as a shell
- *      redirect, so it is absent from argv and `pkill -f <logfile>` matches
- *      nothing (the process survived, verified live).
- *   2. The PORT cannot identify it either, because we deliberately launch with
- *      `--port 0` (the remote picks a free port, and the chosen number is only
- *      known afterwards). A `--port 0` substring match therefore never fires.
- *
- * So the recorded PID is the handle, and the kill targets its process GROUP
- * (`kill -- -PID`): the launch goes through `setsid`, making the child a
- * session/group leader whose PGID equals its PID, so the group contains exactly
- * the process tree we started — never a sibling attach session and never the
- * user's own DSH. Before signalling, the command re-verifies that the PID still
- * exists and still looks like a `dsh` process, so a recycled PID belonging to an
- * unrelated program is not killed.
- *
- * @param {number} pid - the PID reported by the launch.
- * @param {string} [logPath] - this session's log file, removed after the kill so
- *   repeated attaches do not accumulate files in the remote home.
- * @returns {string} a POSIX shell command (always exits 0, never throws).
- */
-export function buildStopCommand(pid, logPath = '') {
-  const p = Number(pid) || 0
-  const cleanup = logPath ? `; rm -f "$HOME/${logPath}" 2>/dev/null` : ''
-  if (p < 1) return `true${cleanup}`
-  return `P=${p}`
-    + `; if [ -d "/proc/$P" ]; then`
-    + ` C=$(tr '\\0' ' ' < "/proc/$P/cmdline" 2>/dev/null);`
-    + ` case "$C" in *dsh*) kill -- -"$P" 2>/dev/null || kill "$P" 2>/dev/null;; esac;`
-    + ` fi${cleanup}; true`
-}
-
-/**
- * One attached remote DSH Web UI: a local listener whose sockets are carried
- * over SSH to the remote's loopback web port.
- *
- * The listener is a plain `net.Server` + `client.forwardOut` pair — the same
- * mechanism `ForwardManager` uses for user-defined forwards. It is implemented
- * here rather than through that manager because this tunnel is bound to a
- * CHOSEN pool (the machine the user picked), whereas `ForwardManager` is wired
- * to the single "active machine" pool; routing an attach through it would send
- * the tunnel to whichever machine happened to be active.
- */
-export class WebAttach {
-  /**
-   * @param {object} opts - injected collaborators (all optional in production).
-   * @param {object} [opts.net] - a `node:net`-compatible module (tests).
-   * @param {number} [opts.localPortStart] - first local port to try.
-   * @param {number} [opts.localPortMax] - how many ports to try before failing.
-   * @param {boolean} [opts.trackPool] - follow the pool across reconnects so a
-   *   dropped SSH connection degrades to "disconnected" and then SELF-HEALS
-   *   instead of being a permanent dead end (issue #54). `true` in production.
-   */
-  constructor(opts = {}) {
-    this.net = opts.net || net
-    // `??`, not `||`: port 0 is a MEANINGFUL value here (let the OS pick), and
-    // `0 || 3088` silently turned "let the OS choose" into a bind on a fixed
-    // port. An explicit 0 now survives, which is also what lets a test ask for a
-    // port without risking a collision with something really listening.
-    this.localPortStart = opts.localPortStart ?? 3088
-    this.localPortMax = opts.localPortMax ?? 50
-    /** Whether the pool's lifecycle hooks are ours to install. */
-    this.trackPool = opts.trackPool !== false
-    /** @type {import('node:net').Server|null} */ this.server = null
-    /** @type {Set<import('node:net').Socket>} */ this.sockets = new Set()
-    /** @type {object|null} */ this.client = null
-    /** The pool this session borrowed its connection from (may be null). */
-    this.pool = null
-    /** Pool hooks we installed, so close() can hand them back untouched. */
-    this.prevOnReady = null
-    this.prevOnCloseHook = null
-    this.hookedPool = null
-    this._installedOnReady = null
-    this._installedOnClose = null
-    this.attachedClient = null
-    /** Bound listener, kept so it can be removed from the client. */
-    this.onClientClose = null
-    this.localPort = 0
-    this.remotePort = 0
-    this.token = ''
-    this.tag = ''
-    this.logPath = ''
-    this.remotePid = 0
-    this.machineId = ''
-    this.command = ''
-    this.startedAt = 0
-    /** Launch knobs remembered so a restart after a drop uses the same ones. */
-    this.waitSeconds = 0
-    this.omittedNoOpen = false
-    /** True while a reconnect-triggered re-resolve is in flight. */
-    this._resolving = false
-    /** True once the borrowed connection has been lost (drives recovery). */
-    this.connectionLost = false
-    /** undefined = "not asked yet"; false = the remote process is provably gone. */
-    this._remoteAlive = undefined
-    /** Set by close(): nothing may re-arm the tunnel afterwards. */
-    this.closed = false
-    /** Set once the borrowed connection has gone away (see _clientLost). */
-    this.lastError = ''
-    /**
-     * A listener that throws EARNS an 'error' event on the emitting object, and
-     * an 'error' with no listener is itself fatal — the very crash this module
-     * exists to prevent. Nothing here is expected to throw, but a lifecycle
-     * callback running on someone else's event loop must never be the reason
-     * the host process dies, so both hooks are wrapped.
-     */
-    this._onReady = (client) => { try { this._attach(client) } catch { /* never fatal */ } }
-    this._onPoolClose = () => { try { this._clientLost('ssh connection closed') } catch { /* never fatal */ } }
-  }
-
-  /** Whether the tunnel is currently listening. */
-  get active() {
-    return this.server !== null
-  }
-
-  /**
-   * Attach to a remote DSH Web UI.
-   *
-   * @param {object} spec - the attach request.
-   * @param {object} spec.pool - the SshPool for the chosen machine.
-   * @param {string} spec.machineId - machine id, for bookkeeping.
-   * @param {string} [spec.command] - remote dsh executable (default `dsh`).
-   * @param {string} [spec.profile] - DSH profile to boot (default `web`).
-   * @param {string} [spec.existingUrl] - attach to an ALREADY running remote
-   *   web UI instead of starting one: its `http://127.0.0.1:<port>/?token=…`.
-   * @param {(cmd: string, opts?: object) => Promise<{stdout: string, code: number}>} [spec.exec]
-   *   - remote exec, defaulting to `pool.exec`.
-   * @returns {Promise<{url: string, localPort: number, remotePort: number}>}
-   */
-  async open(spec) {
-    const {
-      pool, machineId = '', command = 'dsh', profile = 'web',
-      existingUrl = '', dshHome = '', waitSeconds = 45,
-    } = spec || {}
-    if (!pool) throw new Error('web-attach: a machine pool is required')
-    this.machineId = String(machineId || '')
-    this.command = command
-    this.waitSeconds = Number(waitSeconds) || 45
-
-    const exec = spec.exec || ((cmd, opts) => pool.exec(cmd, opts))
-    // Connecting first also surfaces credential/host-key problems BEFORE we
-    // start a remote process we would then have to clean up. `_attach` (rather
-    // than a bare assignment) is what subscribes to the connection's 'close',
-    // so the loss is noticed the moment it happens instead of on the next
-    // request — which is when the old code let ssh2 throw 'Not connected' from
-    // inside a socket callback and take the host process with it (issue #54).
-    this._attach(await pool.connect())
-    this._followPool(pool)
-
-    if (existingUrl) {
-      const parsed = parseWebStart(`dsh web: ${existingUrl}`)
-      if (!parsed) {
-        throw new Error('web-attach: existing URL must look like http://127.0.0.1:<port>/?token=<token>')
-      }
-      this.remotePort = parsed.port
-      this.token = parsed.token
-    } else {
-      this.tag = randomTag()
-      const logPath = `${LOG_NAME}-${this.tag}.log`
-      this.logPath = logPath
-      const timeoutMs = Math.max(30000, (waitSeconds + 20) * 1000)
-      let out = String(await this._launch(exec, { command, profile, logPath, dshHome, waitSeconds, timeoutMs }, false))
-      // Older DSH (0.1.0-rc.6) has no `--no-open` and exits on the unknown flag
-      // before doing anything. Retry once without it: `--no-open` is a nicety
-      // (it stops the REMOTE from opening its own browser), not a requirement,
-      // and a bounded single retry is safe because the first attempt provably
-      // never started a server.
-      if (launchFailed(out) && /unknown option/i.test(out) && /--no-open/.test(out)) {
-        out = String(await this._launch(exec, { command, profile, logPath, dshHome, waitSeconds, timeoutMs }, true))
-        this.omittedNoOpen = true
-      }
-      this.remotePid = parseRemotePid(out)
-      const tail = out.split(TAIL_MARKER).slice(1).join(TAIL_MARKER).trim()
-      // Both failure shapes carry the same diagnostics: the process either died
-      // (its output is the whole story) or is still booting after the deadline.
-      if (launchFailed(out)) {
-        // The remote process is already gone — nothing to clean up.
-        throw new Error('web-attach: the remote DSH exited during startup.'
-          + explainLaunchFailure(tail || out)
-          + ` ${tail || 'no output'}`)
-      }
-      const parsed = parseWebStart(out)
-      if (!parsed) {
-        // A timed-out launch may STILL be booting: the deadline expiring is not
-        // evidence the process died. Kill it before throwing, or every retry
-        // leaks another remote server (measured: three leftover `dsh --port 0`
-        // processes after three failed attaches). A PID of 0 means the launcher
-        // could not report one — nothing to target.
-        if (this.remotePid > 0) {
-          try {
-            await exec(buildStopCommand(this.remotePid, logPath), { timeoutMs: 20000 })
-            this.remotePid = 0
-          } catch { /* the process may already be gone; a leftover is worse than a failed kill attempt only if it repeats, and the next attach replaces this session */ }
-        }
-        throw new Error(
-          `web-attach: the remote DSH did not report a startup token within ${waitSeconds}s.`
-          + ` Check that \`${command}\` exists on the remote and can boot a web profile.`
-          + explainLaunchFailure(tail || out)
-          + (tail ? ` Remote output: ${tail}` : ''),
-        )
-      }
-      this.remotePort = parsed.port
-      this.token = parsed.token
-    }
-
-    await this._listen()
-    this.startedAt = Date.now()
-    return this.describe()
-  }
-
-  /**
-   * Run one launch attempt and return its raw output.
-   *
-   * @param {(cmd: string, opts?: object) => Promise<{stdout: string}>} exec - remote exec.
-   * @param {object} d - launch parameters plus the exec timeout.
-   * @param {boolean} omitNoOpen - drop `--no-open` for old DSH.
-   * @returns {Promise<string>} the launcher's stdout.
-   */
-  async _launch(exec, d, omitNoOpen) {
-    const cmd = buildLaunchCommand({ ...d, omitNoOpen })
-    const res = await exec(cmd, { timeoutMs: d.timeoutMs })
-    return String(res && res.stdout ? res.stdout : '')
-  }
-
-  /** Bind the local listener, advancing past ports already in use. */
-  async _listen() {
-    let lastError = null
-    for (let i = 0; i < this.localPortMax; i++) {
-      const port = this.localPortStart + i
-      try {
-        await this._listenOn(port)
-        this.localPort = port
-        return
-      } catch (err) {
-        lastError = err
-        // Only a taken port is worth retrying; anything else (permissions,
-        // an unusable interface) will not be fixed by the next number.
-        if (err && err.code !== 'EADDRINUSE') break
-      }
-    }
-    throw new Error(
-      `web-attach: no free local port in ${this.localPortStart}..${this.localPortStart + this.localPortMax - 1}`
-      + (lastError ? ` (${lastError.code || lastError.message})` : ''),
-    )
-  }
-
-  _listenOn(port) {
-    return new Promise((resolve, reject) => {
-      const server = this.net.createServer((socket) => this._pipe(socket))
-      const onError = (err) => {
-        server.removeListener('listening', onListening)
-        try { server.close() } catch { /* never listened */ }
-        reject(err)
-      }
-      const onListening = () => {
-        server.removeListener('error', onError)
-        // A post-bind error must not crash the host process: report and drop
-        // the tunnel so the UI can show "disconnected" instead of dying.
-        server.on('error', (err) => {
-          this.server = null
-          this.lastError = String((err && err.message) || err)
-        })
-        this.server = server
-        resolve()
-      }
-      server.once('error', onError)
-      server.once('listening', onListening)
-      // Loopback only: this endpoint reaches a full remote harness, so it must
-      // never be reachable from the network even on a LAN-facing machine.
-      server.listen(port, '127.0.0.1')
-    })
-  }
-
-  /**
-   * Borrow the given connection AND subscribe to its lifecycle.
-   *
-   * Idempotent for the same client (open() calls it once, the pool's `onReady`
-   * calls it again on every reconnect), and it always drops the previous
-   * client's listener first, so one session never accumulates listeners.
-   *
-   * @param {object|null} client - the live ssh2 client / OpenSshConnection.
-   */
-  _attach(client) {
-    if (!client) return
-    if (this.attachedClient && this.attachedClient !== client && this.onClientClose) {
-      // A different client replaced the one we borrowed (pool reconnect). Stop
-      // listening to the old one; its own 'close' has either fired already or
-      // belongs to a connection nobody uses any more.
-      try { this.attachedClient.removeListener('close', this.onClientClose) } catch { /* already gone */ }
-    }
-    this.client = client
-    if (this.attachedClient === client) return
-    this.attachedClient = client
-    this.onClientClose = () => { this._clientLost('ssh connection closed') }
-    // Without this, the object keeps handing out a connection that is already
-    // closed, and `forwardOut` then throws SYNCHRONOUSLY (issue #54).
-    try { client.on('close', this.onClientClose) } catch { /* not an emitter: nothing to watch */ }
-  }
-
-  /** The borrowed connection is gone: keep the session, drop the dead handle. */
-  _clientLost(reason) {
-    this.client = null
-    this.connectionLost = true
-    this.lastError = String(reason || 'ssh connection closed')
-  }
-
-  /**
-   * Register lifecycle hooks on the chosen pool so this tunnel follows it.
-   *
-   * The pool is the one place that both NOTICES a dead connection and later
-   * obtains a fresh one, and it exposes exactly two hooks for that:
-   *   • `onCloseHook` — fires from `close()` and `invalidate()` (the latter is
-   *     the "channel open failed, the pooled connection died server-side"
-   *     self-heal path, which is the likeliest trigger);
-   *   • `onReady` — fires after a successful (re)connect.
-   *
-   * Both are SINGLE-SLOT properties that the plugin already uses for
-   * `ForwardManager` (index.js), so this session CHAINS them rather than
-   * clobbering the forwards manager, and `_detachPool()` restores exactly what
-   * was there — the plugin may legitimately have none, one, or both.
-   *
-   * A real reconnect is the only way the remote port can move, so once the pool
-   * reports a new connection we must re-resolve (token, port) from the remote
-   * log. `onReady` itself cannot await: the pool fires it with the client in
-   * hand, so the re-resolve is kicked off and its failure is recorded rather
-   * than thrown into the pool's callback.
-   *
-   * @param {object} pool - the pool this session attached through.
-   */
-  _followPool(pool) {
-    if (!this.trackPool || this.closed || !pool || this.hookedPool === pool) return
-    this._detachPool()
-    this.pool = pool
-    this.hookedPool = pool
-    const prevReady = typeof pool.onReady === 'function' ? pool.onReady : null
-    const prevClose = typeof pool.onCloseHook === 'function' ? pool.onCloseHook : null
-    this.prevOnReady = prevReady
-    this.prevOnCloseHook = prevClose
-    const installedReady = (client) => {
-      this._attach(client)
-      if (prevReady) { try { prevReady(client) } catch { /* the forwards manager's problem, not ours */ } }
-      // A reconnect lands on a NEW remote process only if the old one died with
-      // the machine; re-read the log to find out which. The re-resolve runs off
-      // the event loop because the pool cannot await it — a hook that throws
-      // would earn an 'error' event on the pool, and an unhandled rejection
-      // would be fatal.
-      if (!this.closed && this.remotePort && !this._resolving) {
-        this._resolving = true
-        Promise.resolve()
-          .then(() => this._refreshRemotePort())
-          .catch((err) => { this.lastError = String((err && err.message) || err) })
-          .then(() => { this._resolving = false })
-      }
-    }
-    const installedClose = () => {
-      this._clientLost('ssh connection closed')
-      if (prevClose) { try { prevClose() } catch { /* never mask our own report */ } }
-    }
-    this._installedOnReady = installedReady
-    this._installedOnClose = installedClose
-    pool.onReady = installedReady
-    pool.onCloseHook = installedClose
-  }
-
-  /**
-   * Hand the pool's hooks back.
-   *
-   * Each slot is restored ONLY when it still holds the wrapper this session
-   * installed. The pool's hooks are single-slot, so if something else has taken
-   * the slot over in the meantime, putting a stale callback back would be worse
-   * than leaving it alone — it would silently disconnect a newer owner.
-   */
-  _detachPool() {
-    const pool = this.hookedPool
-    const mineReady = this._installedOnReady
-    const mineClose = this._installedOnClose
-    this.hookedPool = null
-    this._installedOnReady = null
-    this._installedOnClose = null
-    if (!pool) return
-    if (mineReady && pool.onReady === mineReady) pool.onReady = this.prevOnReady
-    if (mineClose && pool.onCloseHook === mineClose) pool.onCloseHook = this.prevOnCloseHook
-    this.prevOnReady = null
-    this.prevOnCloseHook = null
-  }
-
-  /**
-   * True when the borrowed connection was LOST and this is a session we can
-   * recover: we launched the remote process ourselves (so we hold its PID and
-   * log) and the tunnel is listening (so clients can actually be served).
-   *
-   * The `connectionLost` flag is deliberately separate from `this.client`: the
-   * pool's `onReady` re-attaches the new client BEFORE the re-resolve runs, so a
-   * check on `!this.client` would be false by then and recovery would never
-   * happen.
-   */
-  get _needsReResolve() {
-    return !!(this.connectionLost && !this.closed && this.logPath && this.remotePid && this.localPort)
-  }
-
-  /**
-   * Re-resolve the remote port after a reconnect, restarting the remote `dsh web`
-   * when it did not survive.
-   *
-   * Reaching a remote process needs a live SSH connection, so this reconnects
-   * through the pool first — the pool reports `onReady` for that same client in
-   * production, and in a test double the callback simply never fires.
-   */
-  async _refreshRemotePort() {
-    if (!this._needsReResolve) return
-    const pool = this.pool
-    if (!pool || (typeof pool.connect !== 'function' && typeof pool.exec !== 'function')) return
-    const exec = (cmd, opts) => (typeof pool.exec === 'function'
-      ? pool.exec(cmd, opts)
-      : Promise.reject(new Error('web-attach: the pool cannot run commands')))
-    if (!this.client) await pool.connect()
-    const match = await this._resolveRemotePort(exec)
-    this.remotePort = match.port
-    if (match.token) this.token = match.token
-    this.connectionLost = false
-    this.lastError = ''
-  }
-
-  /**
-   * Read this session's remote log and decide which port is live.
-   *
-   * The startup line is the ONLY source of the port (the launch deliberately
-   * uses `--port 0`), so after the remote process died and was restarted the log
-   * holds BOTH lines — the dead one and the new one — and neither position
-   * identifies the live one. Evidence decides:
-   *   • a process that is still alive on the remote is probed first, and when
-   *     the port it is LISTENING on is one of the logged ports, that is the
-   *     answer (the `dsh web: …` line names its own port, so this is a
-   *     confirmation rather than a second guess);
-   *   • otherwise the newest logged port wins, which is right in every
-   *     single-process case and is the best available answer while the remote is
-   *     still booting (the next reconnect retries).
-   *
-   * @param {(cmd: string, opts?: object) => Promise<{stdout: string}>} exec - remote exec.
-   * @returns {Promise<{port: number, token: string, restarted: boolean}>}
-   */
-  async _resolveRemotePort(exec) {
-    this._remoteAlive = await this._remoteProcessAlive(exec)
-    if (!this._remoteAlive) return this._restartRemote(exec)
-
-    let out = ''
-    try {
-      const res = await exec(`cat "$HOME/${this.logPath}" 2>/dev/null`, { timeoutMs: 15000 })
-      out = String((res && res.stdout) || '')
-    } catch { /* an unreadable log: the listening probe below is the fallback */ }
-    const lines = [...out.matchAll(/dsh web:\s*(http:\/\/[^\s]+)/g)]
-      .map((m) => parseWebStart(`dsh web: ${m[1]}`))
-      .filter(Boolean)
-
-    // What the recorded process actually listens on beats any guess from the
-    // log, because a killed-and-restarted remote leaves BOTH startup lines in
-    // the log and neither position says which one is live.
-    const listening = await this._listeningRemotePort(exec)
-    if (listening) {
-      const match = lines.find((l) => l.port === listening)
-      return { port: listening, token: (match && match.token) || this.token, restarted: false }
-    }
-    if (lines.length) {
-      const newest = lines[lines.length - 1]
-      return { port: newest.port, token: newest.token, restarted: false }
-    }
-    // Alive but silent and not listening yet: the remote is booting. Keep the
-    // last known port and let the next reconnect re-resolve.
-    return { port: this.remotePort, token: this.token, restarted: false }
-  }
-
-  /**
-   * Whether the remote process THIS session started still exists.
-   *
-   * Verified through `/proc`, which is Linux-only; a host that cannot answer (or
-   * a failed probe) is treated as ALIVE on purpose. Restarting a server that was
-   * never dead would spawn a second one and leak a port, whereas assuming life
-   * merely retries a port that turns out to be stale.
-   *
-   * @param {(cmd: string, opts?: object) => Promise<{stdout: string}>} exec - remote exec.
-   * @returns {Promise<boolean>}
-   */
-  async _remoteProcessAlive(exec) {
-    const pid = Number(this.remotePid) || 0
-    if (pid < 1) return true
-    try {
-      const res = await exec(
-        `if [ -d /proc ]; then { [ -d /proc/${pid} ] && echo ALIVE; } || echo GONE; else echo ALIVE; fi`,
-        { timeoutMs: 15000 },
-      )
-      return !/GONE/.test(String((res && res.stdout) || ''))
-    } catch { return true }
-  }
-
-  /**
-   * Ask the remote which loopback port the recorded PID listens on (0 = none).
-   *
-   * Two `/proc` reads are needed because a TCP listener appears in a socket
-   * table (`/proc/net/tcp{,6}`), not in that process's link list: the fd
-   * symlinks give the socket INODE, and the table maps that inode to a port.
-   * Matching on the inode rather than the port is what makes this an
-   * observation instead of a second guess.
-   */
-  async _listeningRemotePort(exec) {
-    const pid = Number(this.remotePid) || 0
-    if (pid < 1) return 0
-    let inodes = ''
-    try {
-      // `ls -l` on a process's own fds needs no privileges for the same uid.
-      const res = await exec(
-        `ls -l /proc/${pid}/fd 2>/dev/null | sed -n 's/.*socket:\\[\\([0-9]*\\)\\].*/\\1/p' | sort -u | tr '\\n' ' '`,
-        { timeoutMs: 15000 },
-      )
-      inodes = String((res && res.stdout) || '').trim()
-    } catch { return 0 }
-    if (!inodes) return 0
-    const want = new Set(inodes.split(/\s+/).filter(Boolean))
-    // `/proc/net/tcp` addresses are little-endian hex, so loopback is
-    // `0100007F` (v4) and `00000000000000000000000001000000` (v6 `::1`).
-    const LOOPBACK = new Set(['0100007F', '00000000000000000000000001000000'])
-    for (const file of ['/proc/net/tcp', '/proc/net/tcp6']) {
-      let text = ''
-      try {
-        const res = await exec(`cat ${file} 2>/dev/null`, { timeoutMs: 15000 })
-        text = String((res && res.stdout) || '')
-      } catch { continue }
-      for (const line of text.split('\n').slice(1)) {
-        const cols = line.trim().split(/\s+/)
-        // sl local_address rem_address st tx:rx tr:tm->when retrnsmt uid timeout inode
-        if (cols.length < 10) continue
-        if (cols[3] !== '0A') continue // TCP_LISTEN
-        if (!want.has(cols[9])) continue
-        const [addr, portHex] = String(cols[1]).split(':')
-        if (!LOOPBACK.has(String(addr).toUpperCase())) continue
-        const port = parseInt(portHex, 16)
-        if (Number.isInteger(port) && port > 0 && port < 65536) return port
-      }
-    }
-    return 0
-  }
-
-  /**
-   * Relaunch the remote `dsh web` because the previous process is gone.
-   *
-   * A restart CANNOT change the listening port locally, and a caller holding a
-   * stale port silently talks to the wrong place at worst and hangs at best, so
-   * the two facts the caller needs are checked: this is a session that launched
-   * its own process (an attach onto someone else's DSH stays hands-off and can
-   * never restart it), and recovery is possible at all (`_refreshRemotePort`
-   * has already returned early when it is not).
-   */
-  async _restartRemote(exec) {
-    const command = this.command || 'dsh'
-    const waitSeconds = this.waitSeconds || 45
-    const timeoutMs = Math.max(30000, (waitSeconds + 20) * 1000)
-    let out = String(await this._launch(exec, { command, profile: 'web', logPath: this.logPath, waitSeconds, timeoutMs }, this.omittedNoOpen === true))
-    if (launchFailed(out) && /unknown option/i.test(out) && /--no-open/.test(out)) {
-      out = String(await this._launch(exec, { command, profile: 'web', logPath: this.logPath, waitSeconds, timeoutMs }, true))
-      this.omittedNoOpen = true
-    }
-    this.remotePid = parseRemotePid(out) || this.remotePid
-    const parsed = parseWebStart(out)
-    if (!parsed) {
-      const tail = out.split(TAIL_MARKER).slice(1).join(TAIL_MARKER).trim()
-      throw new Error('web-attach: the remote DSH could not be restarted after the SSH'
-        + ' connection dropped.' + explainLaunchFailure(tail || out) + (tail ? ` Remote output: ${tail}` : ''))
-    }
-    return { port: parsed.port, token: parsed.token, restarted: true }
-  }
-
-  /** Carry one local socket to the remote web port over SSH. */
-  _pipe(socket) {
-    if (!this.client) { try { socket.destroy() } catch { /* already gone */ } return }
-    // ssh2's forwardOut does NOT report a dead connection through its callback:
-    // when the socket is gone it THROWS 'Not connected' synchronously
-    // (ssh2/lib/client.js: `if (!this._sock || !isWritable(this._sock)) throw`)
-    // and the callback below is never reached — so `if (err)` cannot guard this
-    // path. Uncaught, that throw travels net.Server → TCP.onconnection and kills
-    // the whole host process (issue #54: every page refresh after the pooled
-    // connection dropped took the session, the browser and the remote orphan
-    // down with it). A dead tunnel is a UI state, not a reason to exit.
-    try {
-      this.client.forwardOut('127.0.0.1', 0, '127.0.0.1', this.remotePort, (err, channel) => {
-        if (err) { try { socket.destroy() } catch { /* already gone */ } return }
-        this.sockets.add(socket)
-        socket.pipe(channel)
-        channel.pipe(socket)
-        const kill = () => {
-          try { socket.destroy() } catch { /* already gone */ }
-          try { channel.close() } catch { /* already closed */ }
-          this.sockets.delete(socket)
-        }
-        socket.on('error', kill)
-        channel.on('error', kill)
-        socket.on('close', kill)
-        channel.on('close', kill)
-      })
-    } catch (err) {
-      this._clientLost(String((err && err.message) || err))
-      try { socket.destroy() } catch { /* already gone */ }
-    }
-  }
-
-  /**
-   * The URL to open, carrying the one-time launch token.
-   *
-   * The token is a credential, so it stays on loopback and is never sent
-   * anywhere; DSH exchanges it for a 30-day signed cookie on the first request
-   * and redirects to a clean `/`.
-   */
-  get url() {
-    return `http://127.0.0.1:${this.localPort}/?token=${this.token}`
-  }
-
-  /** The clean URL (no token) — what the address bar shows after the exchange. */
-  get cleanUrl() {
-    return `http://127.0.0.1:${this.localPort}/`
-  }
-
-  /** Serializable status for the settings UI. */
-  describe() {
-    return {
-      machineId: this.machineId,
-      url: this.url,
-      cleanUrl: this.cleanUrl,
-      token: this.token,
-      localPort: this.localPort,
-      remotePort: this.remotePort,
-      active: this.active,
-      startedAt: this.startedAt,
-      // Only a process we launched can be stopped by us; an attach onto the
-      // user's own running instance is deliberately hands-off.
-      managed: this.remotePid > 0,
-      error: this.lastError || '',
-    }
-  }
-
-  /**
-   * Tear the tunnel down. The remote process is left running by default: the
-   * user may be using that DSH elsewhere, and silently killing a process we did
-   * not start is worse than a leftover one. `stopRemote` opts in.
-   *
-   * @param {object} [opts] - close options.
-   * @param {boolean} [opts.stopRemote] - also terminate the remote process.
-   * @param {(cmd: string) => Promise<unknown>} [opts.exec] - remote exec.
-   */
-  async close({ stopRemote = false, exec } = {}) {
-    // Stop following the pool BEFORE any await: a reconnect landing mid-teardown
-    // must not re-arm the tunnel this call is in the middle of closing.
-    this._detachPool()
-    if (this.attachedClient && this.onClientClose) {
-      try { this.attachedClient.removeListener('close', this.onClientClose) } catch { /* already gone */ }
-    }
-    this.attachedClient = null
-    this.onClientClose = null
-    this.pool = null
-    for (const socket of [...this.sockets]) {
-      try { socket.destroy() } catch { /* already gone */ }
-    }
-    this.sockets.clear()
-    const server = this.server
-    this.server = null
-    if (server) await new Promise((resolve) => server.close(() => resolve()))
-    // Only a process we started has a recorded PID; an attach to the user's own
-    // running instance must never signal anything.
-    if (stopRemote && exec && this.remotePid) {
-      await Promise.resolve(exec(buildStopCommand(this.remotePid, this.logPath))).catch(() => {})
-    }
-    this.client = null
-    // The recorded PID and log path are FACTS about this session (the UI shows
-    // them, and `stopRemote` addresses them), so they survive; `closed` is what
-    // stops recovery work from being re-armed after teardown.
-    this.closed = true
-  }
-}
+// dsh-remote — "remote DSH web UI" attach sessions (issue #46).
+//
+// The user's intent behind that issue is "use the local DSH as a CLIENT for a
+// DSH running on another machine": they do NOT want the remote to expose an
+// inbound listener. DSH makes that impossible anyway — `dsh-web-app` hard-
+// rejects `--host 0.0.0.0` (startup.js) because it would put remote code
+// execution on the network, and `dsh-host-webserver` only accepts the
+// 127.0.0.1/0.0.0.0 literals.
+//
+// So this module inverts the direction, exactly like the rest of the plugin:
+// we SSH OUT, start (or reuse) a loopback-bound `dsh web` on the remote, and
+// carry its socket back over our own SSH connection with a local TCP forward.
+// The browser then talks to 127.0.0.1:<localPort> on this machine.
+//
+// Verified end-to-end on a real Linux host (dsh 0.1.5-rc.2): a bare TCP forward
+// is sufficient — the SPA, its assets, AND the `/api/remote.mux` WebSocket
+// upgrade all transit it (HTTP/1.1 101), with no reverse proxy and no host
+// rewriting. That works because DSH binds its auth cookie to the request
+// authority, so a cookie minted against `127.0.0.1:<localPort>` is exactly what
+// the browser then sends back to that same authority.
+//
+// Three details drive the design and are worth stating once:
+//
+//  1. The launch token CANNOT be read from anywhere but stdout. It is a
+//     `randomBytes` value held in a process-local WeakMap and printed on one
+//     line (`dsh web: http://127.0.0.1:<port>/?token=…`). No file, no env var.
+//     Therefore the token is captured by redirecting the remote process's
+//     output to a log file and parsing that line out of it.
+//  2. `--port 0` lets the remote OS pick a free port, and the same printed line
+//     reports which port it chose — so we never have to guess or probe for a
+//     free port on the remote.
+//  3. The pairing that must survive is (token, port) from ONE launch. Reading
+//     them from one line is what keeps them consistent.
+//
+// Everything here is read-only with respect to the remote filesystem except
+// the log file it creates under the remote `$HOME`.
+import net from 'node:net'
+
+/** Remote log file per attach session; `$HOME` is expanded by the remote shell. */
+const LOG_NAME = '.dsh-remote-web'
+
+/**
+ * Parse the `dsh web:` startup line out of a remote log.
+ *
+ * The line looks like:
+ *   dsh web: http://127.0.0.1:43129/?token=Iy7FKuPSI5qJKJUFDyah4B6i8Ul6_hvspiJhiWufYt8
+ * and may be followed by a LAN suffix when the remote bound 0.0.0.0 (it never
+ * will here — we never pass `--host` — but tolerate it rather than break).
+ *
+ * @param {string} text - the remote log contents.
+ * @returns {{url: string, port: number, token: string}|null} parsed facts, or
+ *   null when the line has not appeared yet (still booting) or is malformed.
+ */
+export function parseWebStart(text) {
+  if (typeof text !== 'string' || text === '') return null
+  // Take the LAST match: a reused log file may hold an earlier session's line,
+  // and the newest one is the one this launch produced.
+  const matches = [...text.matchAll(/dsh web:\s*(http:\/\/[^\s]+)/g)]
+  if (matches.length === 0) return null
+  const url = matches[matches.length - 1][1]
+  let parsed
+  try {
+    parsed = new URL(url)
+  } catch {
+    return null
+  }
+  const port = Number(parsed.port)
+  const token = parsed.searchParams.get('token') || ''
+  if (!Number.isInteger(port) || port < 1 || port > 65535) return null
+  if (!token) return null
+  return { url, port, token }
+}
+
+/** True when the log proves the remote process died instead of booting. */
+export function launchFailed(text) {
+  return typeof text === 'string' && text.includes(DEAD_MARKER)
+}
+
+/** Marker the remote launcher echoes when the process exits before printing. */
+const DEAD_MARKER = '__DSH_REMOTE_DIED__'
+/** Marker separating the parsed line from the raw log tail for diagnostics. */
+export const TAIL_MARKER = '__DSH_TAIL__'
+/** Marker carrying the remote PID so the session can stop its own process. */
+export const PID_MARKER = '__DSH_PID__='
+/** Random ASCII tag so two attach sessions never share a remote log file. */
+function randomTag() {
+  return Math.random().toString(36).slice(2, 8) + Date.now().toString(36).slice(-4)
+}
+
+/**
+ * Build the one-shot remote command that launches `dsh web` detached and waits
+ * for its startup line.
+ *
+ * `setsid` + `nohup` + redirected stdio are all required: the SSH channel closes
+ * as soon as this shell exits, and without detaching the SIGHUP would kill the
+ * server we just started. The loop is bounded so a wedged remote cannot hang the
+ * call past the caller's own timeout, and it reports either the parsed line or
+ * the log tail — never silence.
+ *
+ * @param {object} spec - launch parameters.
+ * @param {boolean} [spec.omitNoOpen] - drop `--no-open`. Older DSH (measured on
+ *   0.1.0-rc.6) has no such flag and exits with "unknown option '--no-open'";
+ *   the caller retries with this set rather than assuming a flag exists.
+ * @returns {string} a POSIX shell command line (no newlines, ASCII only).
+ */
+export function buildLaunchCommand({ command, profile, logPath, remotePort = 0, dshHome = '', waitSeconds = 45, omitNoOpen = false }) {
+  const env = dshHome ? `DSH_HOME="${dshHome}" ` : ''
+  const noOpen = omitNoOpen ? '' : ' --no-open'
+  const args = `--profile ${profile} --port ${Number(remotePort) || 0}${noOpen}`
+  const limit = Number(waitSeconds) || 45
+  // The loop is ONE shell element: `do` is a keyword and `do;` is a syntax error
+  // in POSIX sh, so it must be glued to the first command of the body rather
+  // than separated like the other statements.
+  const loop = `while [ $i -lt ${limit} ]; do grep -q 'token=' "$LOG" 2>/dev/null && break`
+    + `; kill -0 $P 2>/dev/null || { echo ${DEAD_MARKER}; break; }`
+    + `; sleep 1; i=$((i+1)); done`
+  return [
+    `LOG="$HOME/${logPath}"`,
+    `: > "$LOG"`,
+    // `command -v setsid` guards hosts without util-linux (macOS); plain nohup
+    // still detaches from the channel's HUP there.
+    `command -v setsid >/dev/null 2>&1 && SET=setsid || SET=`,
+    `${env}$SET nohup ${command} ${args} >"$LOG" 2>&1 </dev/null & P=$!`,
+    `i=0`,
+    loop,
+    `echo "${PID_MARKER}$P"`,
+    `grep -m1 'dsh web:' "$LOG" 2>/dev/null`,
+    `echo ${TAIL_MARKER}`,
+    `tail -20 "$LOG" 2>/dev/null`,
+  ].join('; ')
+}
+
+/** Extract the remote PID from launcher output (0 when absent). */
+export function parseRemotePid(output) {
+  const m = /__DSH_PID__=(\d+)/.exec(String(output || ''))
+  return m ? Number(m[1]) : 0
+}
+
+/**
+ * Recognize a known remote-side reason the web surface could not start, and
+ * return an actionable sentence for it.
+ *
+ * The point is to keep an operator from reading "the remote DSH did not report a
+ * startup token" and concluding the *plugin* is broken. Measured cause: dsh
+ * ≤ 0.1.0-rc.6 depends on `node-pty@1.1.0`, whose published tarball ships
+ * darwin/win32 prebuilds but **no `linux-x64`**, so `dsh web` cannot boot on
+ * Linux at all — it dies loading the native module. 0.1.5-rc.2 moved to
+ * `node-pty@1.2.0-beta.15`, which does ship Linux builds.
+ *
+ * @param {string} output - the remote log tail.
+ * @returns {string} a hint, or '' when nothing is recognized.
+ */
+export function explainLaunchFailure(output) {
+  const text = String(output || '')
+  if (/pty\.node|Failed to load native module/i.test(text)) {
+    return ' The remote could not load node-pty\'s native module. Known cause: dsh <= 0.1.0-rc.6'
+      + ' depends on node-pty@1.1.0, which publishes no linux-x64 prebuild, so `dsh web` cannot start'
+      + ' on Linux at all. Upgrade the remote dsh to 0.1.5-rc.2 or newer (node-pty@1.2.0-beta.15 ships'
+      + ' the Linux build), or point `webAttachCommand` at a newer dsh installed elsewhere.'
+  }
+  if (/is intentionally not supported|--host 0\.0\.0\.0/i.test(text)) {
+    return ' The remote dsh refused an all-interfaces bind. Attach deliberately never passes `--host`;'
+      + ' check `webAttachCommand` does not add one.'
+  }
+  if (/command not found|No such file or directory|not found/i.test(text)) {
+    return ' The remote shell could not run the dsh command. If the remote dsh is not on the SSH login'
+      + ' PATH (a version manager or private prefix), set `webAttachCommand` to its absolute path.'
+  }
+  return ''
+}
+
+/**
+ * Build the command that stops the remote process THIS session started.
+ *
+ * Two things make this less obvious than it looks, both found by running it
+ * against a real host:
+ *   1. The log path cannot identify the process — it appears only as a shell
+ *      redirect, so it is absent from argv and `pkill -f <logfile>` matches
+ *      nothing (the process survived, verified live).
+ *   2. The PORT cannot identify it either, because we deliberately launch with
+ *      `--port 0` (the remote picks a free port, and the chosen number is only
+ *      known afterwards). A `--port 0` substring match therefore never fires.
+ *
+ * So the recorded PID is the handle, and the kill targets its process GROUP
+ * (`kill -- -PID`): the launch goes through `setsid`, making the child a
+ * session/group leader whose PGID equals its PID, so the group contains exactly
+ * the process tree we started — never a sibling attach session and never the
+ * user's own DSH. Before signalling, the command re-verifies that the PID still
+ * exists and still looks like a `dsh` process, so a recycled PID belonging to an
+ * unrelated program is not killed.
+ *
+ * @param {number} pid - the PID reported by the launch.
+ * @param {string} [logPath] - this session's log file, removed after the kill so
+ *   repeated attaches do not accumulate files in the remote home.
+ * @returns {string} a POSIX shell command (always exits 0, never throws).
+ */
+export function buildStopCommand(pid, logPath = '') {
+  const p = Number(pid) || 0
+  const cleanup = logPath ? `; rm -f "$HOME/${logPath}" 2>/dev/null` : ''
+  if (p < 1) return `true${cleanup}`
+  return `P=${p}`
+    + `; if [ -d "/proc/$P" ]; then`
+    + ` C=$(tr '\\0' ' ' < "/proc/$P/cmdline" 2>/dev/null);`
+    + ` case "$C" in *dsh*) kill -- -"$P" 2>/dev/null || kill "$P" 2>/dev/null;; esac;`
+    + ` fi${cleanup}; true`
+}
+
+/**
+ * One attached remote DSH Web UI: a local listener whose sockets are carried
+ * over SSH to the remote's loopback web port.
+ *
+ * The listener is a plain `net.Server` + `client.forwardOut` pair — the same
+ * mechanism `ForwardManager` uses for user-defined forwards. It is implemented
+ * here rather than through that manager because this tunnel is bound to a
+ * CHOSEN pool (the machine the user picked), whereas `ForwardManager` is wired
+ * to the single "active machine" pool; routing an attach through it would send
+ * the tunnel to whichever machine happened to be active.
+ */
+export class WebAttach {
+  /**
+   * @param {object} opts - injected collaborators (all optional in production).
+   * @param {object} [opts.net] - a `node:net`-compatible module (tests).
+   * @param {number} [opts.localPortStart] - first local port to try.
+   * @param {number} [opts.localPortMax] - how many ports to try before failing.
+   * @param {boolean} [opts.trackPool] - follow the pool across reconnects so a
+   *   dropped SSH connection degrades to "disconnected" and then SELF-HEALS
+   *   instead of being a permanent dead end (issue #54). `true` in production.
+   */
+  constructor(opts = {}) {
+    this.net = opts.net || net
+    // `??`, not `||`: port 0 is a MEANINGFUL value here (let the OS pick), and
+    // `0 || 3088` silently turned "let the OS choose" into a bind on a fixed
+    // port. An explicit 0 now survives, which is also what lets a test ask for a
+    // port without risking a collision with something really listening.
+    this.localPortStart = opts.localPortStart ?? 3088
+    this.localPortMax = opts.localPortMax ?? 50
+    /** Whether the pool's lifecycle hooks are ours to install. */
+    this.trackPool = opts.trackPool !== false
+    /** @type {import('node:net').Server|null} */ this.server = null
+    /** @type {Set<import('node:net').Socket>} */ this.sockets = new Set()
+    /** @type {object|null} */ this.client = null
+    /** The pool this session borrowed its connection from (may be null). */
+    this.pool = null
+    /** Pool hooks we installed, so close() can hand them back untouched. */
+    this.prevOnReady = null
+    this.prevOnCloseHook = null
+    this.hookedPool = null
+    this._installedOnReady = null
+    this._installedOnClose = null
+    this.attachedClient = null
+    /** Bound listener, kept so it can be removed from the client. */
+    this.onClientClose = null
+    this.localPort = 0
+    this.remotePort = 0
+    this.token = ''
+    this.tag = ''
+    this.logPath = ''
+    this.remotePid = 0
+    this.machineId = ''
+    this.command = ''
+    this.startedAt = 0
+    /** Launch knobs remembered so a restart after a drop uses the same ones. */
+    this.waitSeconds = 0
+    this.omittedNoOpen = false
+    /** True while a reconnect-triggered re-resolve is in flight. */
+    this._resolving = false
+    /** True once the borrowed connection has been lost (drives recovery). */
+    this.connectionLost = false
+    /** undefined = "not asked yet"; false = the remote process is provably gone. */
+    this._remoteAlive = undefined
+    /** Set by close(): nothing may re-arm the tunnel afterwards. */
+    this.closed = false
+    /** Set once the borrowed connection has gone away (see _clientLost). */
+    this.lastError = ''
+    /**
+     * A listener that throws EARNS an 'error' event on the emitting object, and
+     * an 'error' with no listener is itself fatal — the very crash this module
+     * exists to prevent. Nothing here is expected to throw, but a lifecycle
+     * callback running on someone else's event loop must never be the reason
+     * the host process dies, so both hooks are wrapped.
+     */
+    this._onReady = (client) => { try { this._attach(client) } catch { /* never fatal */ } }
+    this._onPoolClose = () => { try { this._clientLost('ssh connection closed') } catch { /* never fatal */ } }
+  }
+
+  /** Whether the tunnel is currently listening. */
+  get active() {
+    return this.server !== null
+  }
+
+  /**
+   * Attach to a remote DSH Web UI.
+   *
+   * @param {object} spec - the attach request.
+   * @param {object} spec.pool - the SshPool for the chosen machine.
+   * @param {string} spec.machineId - machine id, for bookkeeping.
+   * @param {string} [spec.command] - remote dsh executable (default `dsh`).
+   * @param {string} [spec.profile] - DSH profile to boot (default `web`).
+   * @param {string} [spec.existingUrl] - attach to an ALREADY running remote
+   *   web UI instead of starting one: its `http://127.0.0.1:<port>/?token=…`.
+   * @param {(cmd: string, opts?: object) => Promise<{stdout: string, code: number}>} [spec.exec]
+   *   - remote exec, defaulting to `pool.exec`.
+   * @returns {Promise<{url: string, localPort: number, remotePort: number}>}
+   */
+  async open(spec) {
+    const {
+      pool, machineId = '', command = 'dsh', profile = 'web',
+      existingUrl = '', dshHome = '', waitSeconds = 45,
+    } = spec || {}
+    if (!pool) throw new Error('web-attach: a machine pool is required')
+    this.machineId = String(machineId || '')
+    this.command = command
+    this.waitSeconds = Number(waitSeconds) || 45
+
+    const exec = spec.exec || ((cmd, opts) => pool.exec(cmd, opts))
+    // Connecting first also surfaces credential/host-key problems BEFORE we
+    // start a remote process we would then have to clean up. `_attach` (rather
+    // than a bare assignment) is what subscribes to the connection's 'close',
+    // so the loss is noticed the moment it happens instead of on the next
+    // request — which is when the old code let ssh2 throw 'Not connected' from
+    // inside a socket callback and take the host process with it (issue #54).
+    this._attach(await pool.connect())
+    this._followPool(pool)
+
+    if (existingUrl) {
+      const parsed = parseWebStart(`dsh web: ${existingUrl}`)
+      if (!parsed) {
+        throw new Error('web-attach: existing URL must look like http://127.0.0.1:<port>/?token=<token>')
+      }
+      this.remotePort = parsed.port
+      this.token = parsed.token
+    } else {
+      this.tag = randomTag()
+      const logPath = `${LOG_NAME}-${this.tag}.log`
+      this.logPath = logPath
+      const timeoutMs = Math.max(30000, (waitSeconds + 20) * 1000)
+      let out = String(await this._launch(exec, { command, profile, logPath, dshHome, waitSeconds, timeoutMs }, false))
+      // Older DSH (0.1.0-rc.6) has no `--no-open` and exits on the unknown flag
+      // before doing anything. Retry once without it: `--no-open` is a nicety
+      // (it stops the REMOTE from opening its own browser), not a requirement,
+      // and a bounded single retry is safe because the first attempt provably
+      // never started a server.
+      if (launchFailed(out) && /unknown option/i.test(out) && /--no-open/.test(out)) {
+        out = String(await this._launch(exec, { command, profile, logPath, dshHome, waitSeconds, timeoutMs }, true))
+        this.omittedNoOpen = true
+      }
+      this.remotePid = parseRemotePid(out)
+      const tail = out.split(TAIL_MARKER).slice(1).join(TAIL_MARKER).trim()
+      // Both failure shapes carry the same diagnostics: the process either died
+      // (its output is the whole story) or is still booting after the deadline.
+      if (launchFailed(out)) {
+        // The remote process is already gone — nothing to clean up.
+        throw new Error('web-attach: the remote DSH exited during startup.'
+          + explainLaunchFailure(tail || out)
+          + ` ${tail || 'no output'}`)
+      }
+      const parsed = parseWebStart(out)
+      if (!parsed) {
+        // A timed-out launch may STILL be booting: the deadline expiring is not
+        // evidence the process died. Kill it before throwing, or every retry
+        // leaks another remote server (measured: three leftover `dsh --port 0`
+        // processes after three failed attaches). A PID of 0 means the launcher
+        // could not report one — nothing to target.
+        if (this.remotePid > 0) {
+          try {
+            await exec(buildStopCommand(this.remotePid, logPath), { timeoutMs: 20000 })
+            this.remotePid = 0
+          } catch { /* the process may already be gone; a leftover is worse than a failed kill attempt only if it repeats, and the next attach replaces this session */ }
+        }
+        throw new Error(
+          `web-attach: the remote DSH did not report a startup token within ${waitSeconds}s.`
+          + ` Check that \`${command}\` exists on the remote and can boot a web profile.`
+          + explainLaunchFailure(tail || out)
+          + (tail ? ` Remote output: ${tail}` : ''),
+        )
+      }
+      this.remotePort = parsed.port
+      this.token = parsed.token
+    }
+
+    await this._listen()
+    this.startedAt = Date.now()
+    return this.describe()
+  }
+
+  /**
+   * Run one launch attempt and return its raw output.
+   *
+   * @param {(cmd: string, opts?: object) => Promise<{stdout: string}>} exec - remote exec.
+   * @param {object} d - launch parameters plus the exec timeout.
+   * @param {boolean} omitNoOpen - drop `--no-open` for old DSH.
+   * @returns {Promise<string>} the launcher's stdout.
+   */
+  async _launch(exec, d, omitNoOpen) {
+    const cmd = buildLaunchCommand({ ...d, omitNoOpen })
+    const res = await exec(cmd, { timeoutMs: d.timeoutMs })
+    return String(res && res.stdout ? res.stdout : '')
+  }
+
+  /** Bind the local listener, advancing past ports already in use. */
+  async _listen() {
+    let lastError = null
+    for (let i = 0; i < this.localPortMax; i++) {
+      const port = this.localPortStart + i
+      try {
+        await this._listenOn(port)
+        this.localPort = port
+        return
+      } catch (err) {
+        lastError = err
+        // Only a taken port is worth retrying; anything else (permissions,
+        // an unusable interface) will not be fixed by the next number.
+        if (err && err.code !== 'EADDRINUSE') break
+      }
+    }
+    throw new Error(
+      `web-attach: no free local port in ${this.localPortStart}..${this.localPortStart + this.localPortMax - 1}`
+      + (lastError ? ` (${lastError.code || lastError.message})` : ''),
+    )
+  }
+
+  _listenOn(port) {
+    return new Promise((resolve, reject) => {
+      const server = this.net.createServer((socket) => this._pipe(socket))
+      const onError = (err) => {
+        server.removeListener('listening', onListening)
+        try { server.close() } catch { /* never listened */ }
+        reject(err)
+      }
+      const onListening = () => {
+        server.removeListener('error', onError)
+        // A post-bind error must not crash the host process: report and drop
+        // the tunnel so the UI can show "disconnected" instead of dying.
+        server.on('error', (err) => {
+          this.server = null
+          this.lastError = String((err && err.message) || err)
+        })
+        this.server = server
+        resolve()
+      }
+      server.once('error', onError)
+      server.once('listening', onListening)
+      // Loopback only: this endpoint reaches a full remote harness, so it must
+      // never be reachable from the network even on a LAN-facing machine.
+      server.listen(port, '127.0.0.1')
+    })
+  }
+
+  /**
+   * Borrow the given connection AND subscribe to its lifecycle.
+   *
+   * Idempotent for the same client (open() calls it once, the pool's `onReady`
+   * calls it again on every reconnect), and it always drops the previous
+   * client's listener first, so one session never accumulates listeners.
+   *
+   * @param {object|null} client - the live ssh2 client / OpenSshConnection.
+   */
+  _attach(client) {
+    if (!client) return
+    if (this.attachedClient && this.attachedClient !== client && this.onClientClose) {
+      // A different client replaced the one we borrowed (pool reconnect). Stop
+      // listening to the old one; its own 'close' has either fired already or
+      // belongs to a connection nobody uses any more.
+      try { this.attachedClient.removeListener('close', this.onClientClose) } catch { /* already gone */ }
+    }
+    this.client = client
+    if (this.attachedClient === client) return
+    this.attachedClient = client
+    this.onClientClose = () => { this._clientLost('ssh connection closed') }
+    // Without this, the object keeps handing out a connection that is already
+    // closed, and `forwardOut` then throws SYNCHRONOUSLY (issue #54).
+    try { client.on('close', this.onClientClose) } catch { /* not an emitter: nothing to watch */ }
+  }
+
+  /** The borrowed connection is gone: keep the session, drop the dead handle. */
+  _clientLost(reason) {
+    this.client = null
+    this.connectionLost = true
+    this.lastError = String(reason || 'ssh connection closed')
+  }
+
+  /**
+   * Register lifecycle hooks on the chosen pool so this tunnel follows it.
+   *
+   * The pool is the one place that both NOTICES a dead connection and later
+   * obtains a fresh one, and it exposes exactly two hooks for that:
+   *   • `onCloseHook` — fires from `close()` and `invalidate()` (the latter is
+   *     the "channel open failed, the pooled connection died server-side"
+   *     self-heal path, which is the likeliest trigger);
+   *   • `onReady` — fires after a successful (re)connect.
+   *
+   * Both are SINGLE-SLOT properties that the plugin already uses for
+   * `ForwardManager` (index.js), so this session CHAINS them rather than
+   * clobbering the forwards manager, and `_detachPool()` restores exactly what
+   * was there — the plugin may legitimately have none, one, or both.
+   *
+   * A real reconnect is the only way the remote port can move, so once the pool
+   * reports a new connection we must re-resolve (token, port) from the remote
+   * log. `onReady` itself cannot await: the pool fires it with the client in
+   * hand, so the re-resolve is kicked off and its failure is recorded rather
+   * than thrown into the pool's callback.
+   *
+   * @param {object} pool - the pool this session attached through.
+   */
+  _followPool(pool) {
+    if (!this.trackPool || this.closed || !pool || this.hookedPool === pool) return
+    this._detachPool()
+    this.pool = pool
+    this.hookedPool = pool
+    const prevReady = typeof pool.onReady === 'function' ? pool.onReady : null
+    const prevClose = typeof pool.onCloseHook === 'function' ? pool.onCloseHook : null
+    this.prevOnReady = prevReady
+    this.prevOnCloseHook = prevClose
+    const installedReady = (client) => {
+      this._attach(client)
+      if (prevReady) { try { prevReady(client) } catch { /* the forwards manager's problem, not ours */ } }
+      // A reconnect lands on a NEW remote process only if the old one died with
+      // the machine; re-read the log to find out which. The re-resolve runs off
+      // the event loop because the pool cannot await it — a hook that throws
+      // would earn an 'error' event on the pool, and an unhandled rejection
+      // would be fatal.
+      if (!this.closed && this.remotePort && !this._resolving) {
+        this._resolving = true
+        Promise.resolve()
+          .then(() => this._refreshRemotePort())
+          .catch((err) => { this.lastError = String((err && err.message) || err) })
+          .then(() => { this._resolving = false })
+      }
+    }
+    const installedClose = () => {
+      this._clientLost('ssh connection closed')
+      if (prevClose) { try { prevClose() } catch { /* never mask our own report */ } }
+    }
+    this._installedOnReady = installedReady
+    this._installedOnClose = installedClose
+    pool.onReady = installedReady
+    pool.onCloseHook = installedClose
+  }
+
+  /**
+   * Hand the pool's hooks back.
+   *
+   * Each slot is restored ONLY when it still holds the wrapper this session
+   * installed. The pool's hooks are single-slot, so if something else has taken
+   * the slot over in the meantime, putting a stale callback back would be worse
+   * than leaving it alone — it would silently disconnect a newer owner.
+   */
+  _detachPool() {
+    const pool = this.hookedPool
+    const mineReady = this._installedOnReady
+    const mineClose = this._installedOnClose
+    this.hookedPool = null
+    this._installedOnReady = null
+    this._installedOnClose = null
+    if (!pool) return
+    if (mineReady && pool.onReady === mineReady) pool.onReady = this.prevOnReady
+    if (mineClose && pool.onCloseHook === mineClose) pool.onCloseHook = this.prevOnCloseHook
+    this.prevOnReady = null
+    this.prevOnCloseHook = null
+  }
+
+  /**
+   * True when the borrowed connection was LOST and this is a session we can
+   * recover: we launched the remote process ourselves (so we hold its PID and
+   * log) and the tunnel is listening (so clients can actually be served).
+   *
+   * The `connectionLost` flag is deliberately separate from `this.client`: the
+   * pool's `onReady` re-attaches the new client BEFORE the re-resolve runs, so a
+   * check on `!this.client` would be false by then and recovery would never
+   * happen.
+   */
+  get _needsReResolve() {
+    return !!(this.connectionLost && !this.closed && this.logPath && this.remotePid && this.localPort)
+  }
+
+  /**
+   * Re-resolve the remote port after a reconnect, restarting the remote `dsh web`
+   * when it did not survive.
+   *
+   * Reaching a remote process needs a live SSH connection, so this reconnects
+   * through the pool first — the pool reports `onReady` for that same client in
+   * production, and in a test double the callback simply never fires.
+   */
+  async _refreshRemotePort() {
+    if (!this._needsReResolve) return
+    const pool = this.pool
+    if (!pool || (typeof pool.connect !== 'function' && typeof pool.exec !== 'function')) return
+    const exec = (cmd, opts) => (typeof pool.exec === 'function'
+      ? pool.exec(cmd, opts)
+      : Promise.reject(new Error('web-attach: the pool cannot run commands')))
+    if (!this.client) await pool.connect()
+    const match = await this._resolveRemotePort(exec)
+    this.remotePort = match.port
+    if (match.token) this.token = match.token
+    this.connectionLost = false
+    this.lastError = ''
+  }
+
+  /**
+   * Read this session's remote log and decide which port is live.
+   *
+   * The startup line is the ONLY source of the port (the launch deliberately
+   * uses `--port 0`), so after the remote process died and was restarted the log
+   * holds BOTH lines — the dead one and the new one — and neither position
+   * identifies the live one. Evidence decides:
+   *   • a process that is still alive on the remote is probed first, and when
+   *     the port it is LISTENING on is one of the logged ports, that is the
+   *     answer (the `dsh web: …` line names its own port, so this is a
+   *     confirmation rather than a second guess);
+   *   • otherwise the newest logged port wins, which is right in every
+   *     single-process case and is the best available answer while the remote is
+   *     still booting (the next reconnect retries).
+   *
+   * @param {(cmd: string, opts?: object) => Promise<{stdout: string}>} exec - remote exec.
+   * @returns {Promise<{port: number, token: string, restarted: boolean}>}
+   */
+  async _resolveRemotePort(exec) {
+    this._remoteAlive = await this._remoteProcessAlive(exec)
+    if (!this._remoteAlive) return this._restartRemote(exec)
+
+    let out = ''
+    try {
+      const res = await exec(`cat "$HOME/${this.logPath}" 2>/dev/null`, { timeoutMs: 15000 })
+      out = String((res && res.stdout) || '')
+    } catch { /* an unreadable log: the listening probe below is the fallback */ }
+    const lines = [...out.matchAll(/dsh web:\s*(http:\/\/[^\s]+)/g)]
+      .map((m) => parseWebStart(`dsh web: ${m[1]}`))
+      .filter(Boolean)
+
+    // What the recorded process actually listens on beats any guess from the
+    // log, because a killed-and-restarted remote leaves BOTH startup lines in
+    // the log and neither position says which one is live.
+    const listening = await this._listeningRemotePort(exec)
+    if (listening) {
+      const match = lines.find((l) => l.port === listening)
+      return { port: listening, token: (match && match.token) || this.token, restarted: false }
+    }
+    if (lines.length) {
+      const newest = lines[lines.length - 1]
+      return { port: newest.port, token: newest.token, restarted: false }
+    }
+    // Alive but silent and not listening yet: the remote is booting. Keep the
+    // last known port and let the next reconnect re-resolve.
+    return { port: this.remotePort, token: this.token, restarted: false }
+  }
+
+  /**
+   * Whether the remote process THIS session started still exists.
+   *
+   * Verified through `/proc`, which is Linux-only; a host that cannot answer (or
+   * a failed probe) is treated as ALIVE on purpose. Restarting a server that was
+   * never dead would spawn a second one and leak a port, whereas assuming life
+   * merely retries a port that turns out to be stale.
+   *
+   * @param {(cmd: string, opts?: object) => Promise<{stdout: string}>} exec - remote exec.
+   * @returns {Promise<boolean>}
+   */
+  async _remoteProcessAlive(exec) {
+    const pid = Number(this.remotePid) || 0
+    if (pid < 1) return true
+    try {
+      const res = await exec(
+        `if [ -d /proc ]; then { [ -d /proc/${pid} ] && echo ALIVE; } || echo GONE; else echo ALIVE; fi`,
+        { timeoutMs: 15000 },
+      )
+      return !/GONE/.test(String((res && res.stdout) || ''))
+    } catch { return true }
+  }
+
+  /**
+   * Ask the remote which loopback port the recorded PID listens on (0 = none).
+   *
+   * Two `/proc` reads are needed because a TCP listener appears in a socket
+   * table (`/proc/net/tcp{,6}`), not in that process's link list: the fd
+   * symlinks give the socket INODE, and the table maps that inode to a port.
+   * Matching on the inode rather than the port is what makes this an
+   * observation instead of a second guess.
+   */
+  async _listeningRemotePort(exec) {
+    const pid = Number(this.remotePid) || 0
+    if (pid < 1) return 0
+    let inodes = ''
+    try {
+      // `ls -l` on a process's own fds needs no privileges for the same uid.
+      const res = await exec(
+        `ls -l /proc/${pid}/fd 2>/dev/null | sed -n 's/.*socket:\\[\\([0-9]*\\)\\].*/\\1/p' | sort -u | tr '\\n' ' '`,
+        { timeoutMs: 15000 },
+      )
+      inodes = String((res && res.stdout) || '').trim()
+    } catch { return 0 }
+    if (!inodes) return 0
+    const want = new Set(inodes.split(/\s+/).filter(Boolean))
+    // `/proc/net/tcp` addresses are little-endian hex, so loopback is
+    // `0100007F` (v4) and `00000000000000000000000001000000` (v6 `::1`).
+    const LOOPBACK = new Set(['0100007F', '00000000000000000000000001000000'])
+    for (const file of ['/proc/net/tcp', '/proc/net/tcp6']) {
+      let text = ''
+      try {
+        const res = await exec(`cat ${file} 2>/dev/null`, { timeoutMs: 15000 })
+        text = String((res && res.stdout) || '')
+      } catch { continue }
+      for (const line of text.split('\n').slice(1)) {
+        const cols = line.trim().split(/\s+/)
+        // sl local_address rem_address st tx:rx tr:tm->when retrnsmt uid timeout inode
+        if (cols.length < 10) continue
+        if (cols[3] !== '0A') continue // TCP_LISTEN
+        if (!want.has(cols[9])) continue
+        const [addr, portHex] = String(cols[1]).split(':')
+        if (!LOOPBACK.has(String(addr).toUpperCase())) continue
+        const port = parseInt(portHex, 16)
+        if (Number.isInteger(port) && port > 0 && port < 65536) return port
+      }
+    }
+    return 0
+  }
+
+  /**
+   * Relaunch the remote `dsh web` because the previous process is gone.
+   *
+   * A restart CANNOT change the listening port locally, and a caller holding a
+   * stale port silently talks to the wrong place at worst and hangs at best, so
+   * the two facts the caller needs are checked: this is a session that launched
+   * its own process (an attach onto someone else's DSH stays hands-off and can
+   * never restart it), and recovery is possible at all (`_refreshRemotePort`
+   * has already returned early when it is not).
+   */
+  async _restartRemote(exec) {
+    const command = this.command || 'dsh'
+    const waitSeconds = this.waitSeconds || 45
+    const timeoutMs = Math.max(30000, (waitSeconds + 20) * 1000)
+    let out = String(await this._launch(exec, { command, profile: 'web', logPath: this.logPath, waitSeconds, timeoutMs }, this.omittedNoOpen === true))
+    if (launchFailed(out) && /unknown option/i.test(out) && /--no-open/.test(out)) {
+      out = String(await this._launch(exec, { command, profile: 'web', logPath: this.logPath, waitSeconds, timeoutMs }, true))
+      this.omittedNoOpen = true
+    }
+    this.remotePid = parseRemotePid(out) || this.remotePid
+    const parsed = parseWebStart(out)
+    if (!parsed) {
+      const tail = out.split(TAIL_MARKER).slice(1).join(TAIL_MARKER).trim()
+      throw new Error('web-attach: the remote DSH could not be restarted after the SSH'
+        + ' connection dropped.' + explainLaunchFailure(tail || out) + (tail ? ` Remote output: ${tail}` : ''))
+    }
+    return { port: parsed.port, token: parsed.token, restarted: true }
+  }
+
+  /** Carry one local socket to the remote web port over SSH. */
+  _pipe(socket) {
+    if (!this.client) { try { socket.destroy() } catch { /* already gone */ } return }
+    // ssh2's forwardOut does NOT report a dead connection through its callback:
+    // when the socket is gone it THROWS 'Not connected' synchronously
+    // (ssh2/lib/client.js: `if (!this._sock || !isWritable(this._sock)) throw`)
+    // and the callback below is never reached — so `if (err)` cannot guard this
+    // path. Uncaught, that throw travels net.Server → TCP.onconnection and kills
+    // the whole host process (issue #54: every page refresh after the pooled
+    // connection dropped took the session, the browser and the remote orphan
+    // down with it). A dead tunnel is a UI state, not a reason to exit.
+    try {
+      this.client.forwardOut('127.0.0.1', 0, '127.0.0.1', this.remotePort, (err, channel) => {
+        if (err) { try { socket.destroy() } catch { /* already gone */ } return }
+        this.sockets.add(socket)
+        socket.pipe(channel)
+        channel.pipe(socket)
+        const kill = () => {
+          try { socket.destroy() } catch { /* already gone */ }
+          try { channel.close() } catch { /* already closed */ }
+          this.sockets.delete(socket)
+        }
+        socket.on('error', kill)
+        channel.on('error', kill)
+        socket.on('close', kill)
+        channel.on('close', kill)
+      })
+    } catch (err) {
+      this._clientLost(String((err && err.message) || err))
+      try { socket.destroy() } catch { /* already gone */ }
+    }
+  }
+
+  /**
+   * The URL to open, carrying the one-time launch token.
+   *
+   * The token is a credential, so it stays on loopback and is never sent
+   * anywhere; DSH exchanges it for a 30-day signed cookie on the first request
+   * and redirects to a clean `/`.
+   */
+  get url() {
+    return `http://127.0.0.1:${this.localPort}/?token=${this.token}`
+  }
+
+  /** The clean URL (no token) — what the address bar shows after the exchange. */
+  get cleanUrl() {
+    return `http://127.0.0.1:${this.localPort}/`
+  }
+
+  /** Serializable status for the settings UI. */
+  describe() {
+    return {
+      machineId: this.machineId,
+      url: this.url,
+      cleanUrl: this.cleanUrl,
+      token: this.token,
+      localPort: this.localPort,
+      remotePort: this.remotePort,
+      active: this.active,
+      startedAt: this.startedAt,
+      // Only a process we launched can be stopped by us; an attach onto the
+      // user's own running instance is deliberately hands-off.
+      managed: this.remotePid > 0,
+      error: this.lastError || '',
+    }
+  }
+
+  /**
+   * Tear the tunnel down. The remote process is left running by default: the
+   * user may be using that DSH elsewhere, and silently killing a process we did
+   * not start is worse than a leftover one. `stopRemote` opts in.
+   *
+   * @param {object} [opts] - close options.
+   * @param {boolean} [opts.stopRemote] - also terminate the remote process.
+   * @param {(cmd: string) => Promise<unknown>} [opts.exec] - remote exec.
+   */
+  async close({ stopRemote = false, exec } = {}) {
+    // Stop following the pool BEFORE any await: a reconnect landing mid-teardown
+    // must not re-arm the tunnel this call is in the middle of closing.
+    this._detachPool()
+    if (this.attachedClient && this.onClientClose) {
+      try { this.attachedClient.removeListener('close', this.onClientClose) } catch { /* already gone */ }
+    }
+    this.attachedClient = null
+    this.onClientClose = null
+    this.pool = null
+    for (const socket of [...this.sockets]) {
+      try { socket.destroy() } catch { /* already gone */ }
+    }
+    this.sockets.clear()
+    const server = this.server
+    this.server = null
+    if (server) await new Promise((resolve) => server.close(() => resolve()))
+    // Only a process we started has a recorded PID; an attach to the user's own
+    // running instance must never signal anything.
+    if (stopRemote && exec && this.remotePid) {
+      await Promise.resolve(exec(buildStopCommand(this.remotePid, this.logPath))).catch(() => {})
+    }
+    this.client = null
+    // The recorded PID and log path are FACTS about this session (the UI shows
+    // them, and `stopRemote` addresses them), so they survive; `closed` is what
+    // stops recovery work from being re-armed after teardown.
+    this.closed = true
+  }
+}