dsh-plugin-rea 2.0.6 → 2.0.8

package/bin/rea-shim.mjs

@@ -1,6 +1,6 @@
 #!/usr/bin/env node
 /**
- * rea-shim — the process the DSH MCP row launches.
+ * rea-shim 鈥?the process the DSH MCP row launches.
  *
  * Why this exists: the published `cordis.patch.yml` must carry no machine path,
  * yet REA needs absolute paths for every dependency. The shim is the seam: it
@@ -10,12 +10,13 @@
  * It downloads nothing, writes nothing, and never touches PATH.
  *
  * Exit codes:  0 child exited normally   * child's own status
- *              78 EX_CONFIG  bundle not provisioned (or unusable) — the row has
+ *              78 EX_CONFIG  bundle not provisioned (or unusable) 鈥?the row has
  *                            `failOnStartupError: false`, so the harness stays up
  *                            and the message is the operator's remediation.
  */
-import { spawn } from "node:child_process";
-import { existsSync, mkdirSync, readFileSync, rmSync, writeFileSync } from "node:fs";
+import { spawn, spawnSync } from "node:child_process";
+import { createHash } from "node:crypto";
+import { existsSync, mkdirSync, readFileSync, rmSync, statSync, writeFileSync } from "node:fs";
 import { homedir } from "node:os";
 import path from "node:path";
 import process from "node:process";
@@ -32,6 +33,36 @@ function dshHome() {
   return path.join(homedir(), ".dsh");
 }
 
+/**
+ * The DSH home recorded by setup, or null.
+ *
+ * Only the mirrored copy needs this: the harness strips `DSH_*` from an MCP
+ * child, so with a custom home the environment cannot answer. Read once per
+ * process — there is one bundle root per process and this is a startup query.
+ */
+let recordedDshHomeCache = null;
+function recordedDshHome(root) {
+  if (recordedDshHomeCache !== null) return recordedDshHomeCache;
+  let value = null;
+  try {
+    const record = JSON.parse(readFileSync(path.join(root, "mirror.json"), "utf8"));
+    if (typeof record?.dshHome === "string" && record.dshHome.trim() !== "" && path.isAbsolute(record.dshHome)) {
+      value = path.resolve(record.dshHome);
+    }
+  } catch {
+    value = null;
+  }
+  recordedDshHomeCache = value;
+  return value;
+}
+
+/** The home to look in when the environment has been scrubbed. */
+function dshHomeFor(root) {
+  const fromEnv = process.env.DSH_HOME;
+  if (fromEnv && fromEnv.trim() !== "") return path.resolve(fromEnv);
+  return recordedDshHome(root) ?? path.join(homedir(), ".dsh");
+}
+
 /**
  * Where setup writes the machine-local truth.
  *
@@ -43,11 +74,11 @@ function dshHome() {
  *    DSH_HOME: the harness strips every `DSH_*` name from an MCP child's
  *    environment (case-insensitively, because Windows is), so a custom DSH_HOME
  *    would send it to `~/.dsh`, where it would report "not provisioned" for a
- *    bundle sitting right next to it — and start a second ~800 MB run there.
+ *    bundle sitting right next to it 鈥?and start a second ~800 MB run there.
  *  * The **checkout** copy (`<plugin>/bin/rea-shim.mjs`) is what a human runs, e.g.
  *    `npm run bundle:status`, from a shell where DSH_HOME *is* visible. Its own
  *    directory is the plugin, not the bundle, so resolving to itself would report
- *    "not provisioned" on a fully provisioned machine — and a bare
+ *    "not provisioned" on a fully provisioned machine 鈥?and a bare
  *    `node bin/rea-shim.mjs mcp` would provision ~800 MB into the checkout.
  *
  * Only the checkout ships `cordis.patch.yml`, so that file is the discriminator.
@@ -80,7 +111,7 @@ function usableSetup(dir) {
  * The mirrored copy is a snapshot: after a plugin upgrade it still holds the
  * previous manifest, so it would provision the previous versions forever. setup
  * writes <bundle>/mirror.json naming the real checkout; prefer that while it is
- * usable, and fall back to the local copy — which is also all a first run has.
+ * usable, and fall back to the local copy 鈥?which is also all a first run has.
  */
 function provisioningScript(root) {
   try {
@@ -104,13 +135,16 @@ function provisioningScript(root) {
  *
  * The parse is deliberately small: the two files it cares about are block
  * mappings with one id/serverName per row, and anything it cannot attribute is
- * ignored — a missed hint is acceptable, a false alarm is not. The mirrored shim
- * cannot see a custom DSH_HOME (the harness strips DSH_* from an MCP child), so
- * this falls back to `~/.dsh` like the rest of the bundle; the row has the same
- * limitation.
+ * ignored — a missed hint is acceptable, a false alarm is not.
+ *
+ * Finding the file is the part that needed care: the harness strips every DSH_*
+ * variable from an MCP child, so a mirrored copy with a custom DSH_HOME used to
+ * read `~/.dsh` — the wrong file, and therefore a conflict it could never report.
+ * `setup.mjs` writes the home it ran under into `mirror.json`, and that record is
+ * what this uses when the environment has nothing to say.
  */
-function foreignReaRegistration() {
-  const patch = path.join(dshHome(), "cordis.patch.yml");
+function foreignReaRegistration(root) {
+  const patch = path.join(dshHomeFor(root), "cordis.patch.yml");
   let text;
   try {
     text = readFileSync(patch, "utf8");
@@ -133,7 +167,7 @@ function foreignReaRegistration() {
 /**
  * The provisioning lock is a small JSON record: { state, pid, at, exit? }.
  * `state: "running"` is written when a run is started (setup's own pid once it
- * exists), and the failing run rewrites it as `state: "failed"` — it is kept on
+ * exists), and the failing run rewrites it as `state: "failed"` 鈥?it is kept on
  * purpose so a broken setup cannot be re-downloaded in a loop. A legacy
  * plain-number lock is still honoured as "running".
  */
@@ -171,7 +205,7 @@ function pidAlive(pid) {
  * a killed attempt is simply continued on the next start.
  *
  * Set REA_BUNDLE_AUTO_PROVISION=0 to require an explicit `setup.mjs` run.
- * Returns { started, reason, logPath } — the caller words its message from
+ * Returns { started, reason, logPath } 鈥?the caller words its message from
  * `reason`, because "provisioning started in the background" was also printed
  * for a lock left behind by a run that had already failed, and for one whose
  * process was long gone.
@@ -208,7 +242,7 @@ function maybeAutoProvision(root) {
     // Reached when the lock is older than the freshness window, or its holder is
     // gone and never wrote an outcome (harness restart, power loss, a kill): what
     // it records is not happening any more. This removal is the whole point of the
-    // freshness rule — without it the exclusive create below failed with EEXIST
+    // freshness rule 鈥?without it the exclusive create below failed with EEXIST
     // *forever*, auto-provisioning never ran again, and main() printed nothing at
     // all for that outcome. The lock's own "honour it only while it is fresh"
     // comment described behaviour the code did not have.
@@ -265,15 +299,20 @@ function fail(message, detail) {
   if (detail) lines.push("", "  detail: " + detail);
   // Point at the real checkout when the record allows it: running the mirrored
   // copy is fine, but it can only ever provision the manifest it was mirrored
-  // with — and a bundle mirrored before the manifest travelled with it has no
+  // with 鈥?and a bundle mirrored before the manifest travelled with it has no
   // usable copy at all, so the honest answer is then "run the plugin's own script".
-  const setupScript = provisioningScript(bundleRoot());
+  const root = bundleRoot();
+  const setupScript = provisioningScript(root);
   lines.push(
     "",
     "  Finish setup with:",
     setupScript
-      ? "    node \"" + setupScript + "\""
-      : "    node \"<installed plugin directory>/scripts/setup.mjs\"",
+      // `--root` is not decoration: a shim started with REA_BUNDLE_ROOT (or with
+      // DSH_HOME unset in a row's scrubbed environment) resolves a root that
+      // setup's own default need not agree with, and a remediation command that
+      // provisions into another root leaves this row failing exactly as before.
+      ? "    node \"" + setupScript + "\" --root \"" + root + "\""
+      : "    node \"<installed plugin directory>/scripts/setup.mjs\" --root \"" + root + "\"",
   );
   if (!setupScript) {
     lines.push("", "  (Run it from the installed plugin directory: the copy in this bundle has no manifest.)");
@@ -311,7 +350,7 @@ function resolveRea(config, root) {
   // The fallback has to match where setup actually installs the package:
   // `npm install --prefix <root>/vendor` resolves into vendor/node_modules, so a
   // fallback without that segment named a directory that never exists. It only
-  // fires when an older bundle.json has no paths.reaEntry — which is exactly when
+  // fires when an older bundle.json has no paths.reaEntry 鈥?which is exactly when
   // a wrong fallback turns a usable install into "the vendored REA CLI is gone".
   const reaEntry = paths.reaEntry
     ?? path.join(root, "vendor", "node_modules", "rea-agents", "scripts", "rea.mjs");
@@ -363,11 +402,540 @@ function childEnvironment(config) {
   return env;
 }
 
+/**
+ * Launch modes. 2.x exists to supply a runtime the machine does not have; when
+ * the machine already has one, the honest thing to do is run it.
+ *
+ * The 1.x line was exactly that: `rea mcp` off PATH, user-managed Node and
+ * Ghidra. So "use the system install" is not a new code path invented here 鈥?it
+ * is the 1.x contract, reachable from the 2.x row. `auto` (the default) prefers
+ * the system install and falls back to the provisioned bundle; `bundle` pins the
+ * old behaviour; `system` insists (and fails loudly instead of provisioning
+ * 800 MB behind an operator who asked for their own REA).
+ */
+const LAUNCH_MODES = ["auto", "system", "bundle"];
+
+/**
+ * How long a candidate may take to answer `--version`.
+ *
+ * Measured on a real npm-global install: `rea --version` through its `.cmd`
+ * shim takes ~2.4 s cold (cmd.exe -> node -> the CLI's own module graph), and
+ * 2.8-3.0 s with four probes in flight. A 3 s budget was therefore a coin flip:
+ * half the runs rejected a perfectly good REA and silently fell back to the
+ * bundle. 8 s leaves room for a slow disk and a loaded machine while still
+ * bounding the one-off cost at startup.
+ */
+const PROBE_TIMEOUT_MS = 8_000;
+
+/**
+ * What the environment says about the launch mode.
+ *
+ * An unknown value is reported and treated as `auto`: the row is launched by
+ * the harness with nobody watching stderr, so refusing to start over a typo
+ * would read as "the plugin is broken". `auto` is the mode that still works.
+ */
+function launchPreference(env = process.env) {
+  const raw = (env.REA_BUNDLE_LAUNCH ?? "").trim().toLowerCase();
+  if (raw === "") return { value: "auto", invalid: null };
+  return LAUNCH_MODES.includes(raw)
+    ? { value: raw, invalid: null }
+    : { value: "auto", invalid: raw };
+}
+
+/** True when `child` is `parent` or lives under it (case-insensitive on Windows). */
+function isInside(child, parent) {
+  const normalise = (value) => (process.platform === "win32" ? value.toLowerCase() : value);
+  const target = normalise(path.resolve(child));
+  const base = normalise(path.resolve(parent));
+  if (target === base) return true;
+  const relative = path.relative(base, target);
+  return relative !== "" && !relative.startsWith("..") && !path.isAbsolute(relative);
+}
+
+/**
+ * Executable names a REA install provides on this platform.
+ *
+ * npm's global install writes a shell script plus `.cmd`/`.ps1` shims on
+ * Windows, and only the `.cmd` one is spawnable without a shell. `rea.exe` is
+ * accepted for a native/standalone install.
+ */
+function systemReaNames(platform = process.platform) {
+  return platform === "win32" ? ["rea.cmd", "rea.exe", "rea.bat"] : ["rea"];
+}
+
+/**
+ * Every path worth probing, most explicit first.
+ *
+ * `REA_BUNDLE_SYSTEM_REA` names one binary outright 鈥?that is both the test
+ * seam and a real operator need (an install that is not on the harness's PATH,
+ * which on Desktop is not the user's shell PATH).
+ */
+function systemReaCandidates({ env = process.env, platform = process.platform } = {}) {
+  const candidates = [];
+  const explicit = (env.REA_BUNDLE_SYSTEM_REA ?? "").trim();
+  if (explicit !== "") candidates.push(path.resolve(explicit));
+  const searchPath = env.PATH ?? env.Path ?? "";
+  for (const raw of searchPath.split(path.delimiter)) {
+    // A quoted PATH entry (`PATH="C:\Program Files\v66";…`) is legal in Windows
+    // environments and `path.join` would treat the quotes as part of the name, so
+    // the install would be invisible. A relative entry is skipped rather than
+    // resolved: its meaning depends on the cwd THIS process happens to have, and
+    // a candidate found that way would be launched by a path that means something
+    // else to whoever reads the status report.
+    const dir = raw.trim().replace(/^"(.*)"$/, "$1");
+    if (dir === "" || !path.isAbsolute(dir)) continue;
+    for (const name of systemReaNames(platform)) candidates.push(path.join(dir, name));
+  }
+  const seen = new Set();
+  return candidates.filter((candidate) => {
+    const key = process.platform === "win32" ? candidate.toLowerCase() : candidate;
+    if (seen.has(key)) return false;
+    seen.add(key);
+    return true;
+  });
+}
+
+/**
+ * Kill a process and its children.
+ *
+ * `spawnSync`'s timeout kills only the process it started — for a `.cmd` shim
+ * that is cmd.exe, and the `node` running inside it keeps the inherited stdio
+ * handles open and keeps working. Taskkill's `/T` is the one portable-enough way
+ * to get the tree on Windows; POSIX uses the process itself (the candidate is a
+ * shebang script there, so there is no intermediate shell).
+ */
+function killTree(pid, platform = process.platform) {
+  if (typeof pid !== "number" || pid <= 0) return;
+  try {
+    if (platform === "win32") {
+      const taskkill = path.join(process.env.SystemRoot ?? process.env.SYSTEMROOT ?? "C:\\Windows", "System32", "taskkill.exe");
+      spawnSync(taskkill, ["/pid", String(pid), "/T", "/F"], {
+        stdio: "ignore",
+        windowsHide: true,
+        timeout: 5_000,
+      });
+      return;
+    }
+    try {
+      process.kill(pid, "SIGKILL");
+    } catch {
+      /* already gone */
+    }
+  } catch {
+    /* best effort: a failed cleanup must not change the verdict */
+  }
+}
+
+/** True when the path is a regular file (a directory named `rea.cmd` is not). */
+function isFileSync(file) {
+  try {
+    return statSync(file).isFile();
+  } catch {
+    return false;
+  }
+}
+
+/** A command line that survives `cmd /c` on Windows: quoted, quotes doubled. */
+function cmdQuote(value) {
+  return `"${String(value).replace(/"/g, '""')}"`;
+}
+
+/**
+ * The shell used to run a `.cmd`/`.bat` candidate.
+ *
+ * Absolute on purpose: `spawn(comspec())` is resolved against the CHILD's PATH,
+ * and an MCP row inherits a scrubbed environment. A machine with a working REA
+ * but a PATH that does not include System32 would then be judged "no REA",
+ * which is precisely the wrong answer for the feature that exists to find it.
+ */
+function comspec() {
+  const fromEnv = process.env.ComSpec ?? process.env.COMSPEC;
+  // A ComSpec that names nothing is a real (if unusual) state, and trusting it
+  // would judge every `.cmd` REA "spawn-failed" — i.e. report the machine as
+  // having no REA because its shell pointer is broken. Fall back to the path
+  // Windows itself would use.
+  if (typeof fromEnv === "string" && fromEnv.trim() !== "" && isFileSync(fromEnv)) return fromEnv;
+  const systemRoot = process.env.SystemRoot ?? process.env.SYSTEMROOT ?? "C:\\Windows";
+  return path.join(systemRoot, "System32", "cmd.exe");
+}
+
+/**
+ * Ask a candidate whether it is a working REA.
+ *
+ * A file existing on PATH is not the same as a working install: npx leftovers,
+ * half-removed globals and a shim whose Node is gone all sit on PATH and all
+ * fail at once. One bounded `--version` separates them, and a slow or hanging
+ * answer counts as "not usable" — the fallback (the bundle) still works, while
+ * delegating to a broken binary would take the tools down with it.
+ *
+ * Two things here are not incidental:
+ *
+ *  * The probe runs with `REA_BUNDLE_DELEGATED=1`. A candidate that is itself a
+ *    wrapper around this shim would otherwise be probed by a nested shim that
+ *    probes the wrapper — measured: a process chain that keeps respawning
+ *    itself. The flag makes the nested run take the bundle path and return.
+ *  * It is asynchronous, not `spawnSync` with a timeout. `spawnSync` reaps the
+ *    child it started, and its timeout kills only that child: for a `.cmd` shim
+ *    that means cmd.exe dies while the REA inside it keeps running and holds
+ *    the pipes. Killing the TREE is only possible while the parent is alive,
+ *    which means owning the wait ourselves.
+ */
+function probeRea(file, { platform = process.platform, timeout = PROBE_TIMEOUT_MS } = {}) {
+  return new Promise((resolve) => {
+    // `existsSync` says nothing about the shape: a DIRECTORY named `rea.cmd`
+    // exists, fails to spawn, and would be reported as a broken install instead
+    // of as a name collision. Only a regular file is a candidate.
+    if (!isFileSync(file)) {
+      resolve({ ok: false, reason: "not a file" });
+      return;
+    }
+    const useShell = platform === "win32" && /\.(cmd|bat)$/i.test(file);
+    // `REA_BUNDLE_AUTO_PROVISION=0` is part of the same guard: if the candidate
+    // turns out to be a wrapper around THIS shim, the nested run must not start
+    // an 800 MB provisioning job just because a probe asked it for a version.
+    const env = { ...process.env, REA_BUNDLE_DELEGATED: "1", REA_BUNDLE_AUTO_PROVISION: "0" };
+    // `cmd /d /s /c "<line>"`: Windows only. Two details are load-bearing.
+    // `/s` strips the first and last quote, so the whole line is quoted; and the
+    // arguments are passed VERBATIM, because Node otherwise backslash-escapes the
+    // quotes it sees and cmd.exe does not understand `\"` — every candidate then
+    // answers "exit 1", which reads as "no REA on this machine".
+    const child = useShell
+      ? spawn(comspec(), ["/d", "/s", "/c", `"${[cmdQuote(file), "--version"].join(" ")}"`], {
+          env,
+          stdio: ["ignore", "pipe", "pipe"],
+          windowsHide: true,
+          windowsVerbatimArguments: true,
+        })
+      : spawn(file, ["--version"], { env, stdio: ["ignore", "pipe", "pipe"], windowsHide: true });
+    let output = "";
+    let settled = false;
+    let timer = null;
+    const finish = (result) => {
+      if (settled) return;
+      settled = true;
+      if (timer !== null) clearTimeout(timer);
+      resolve(result);
+    };
+    timer = setTimeout(() => {
+      killTree(child.pid, platform);
+      finish({ ok: false, reason: "timeout" });
+    }, timeout);
+    child.stdout.on("data", (chunk) => { output += String(chunk); });
+    child.stderr.on("data", (chunk) => { output += String(chunk); });
+    child.on("error", () => finish({ ok: false, reason: "spawn-failed" }));
+    child.on("close", (code) => {
+      if (code !== 0) {
+        finish({ ok: false, reason: `exit ${String(code)}` });
+        return;
+      }
+      // A REA install prints its version. Anything else (a shim that swallowed
+      // the flag, an unrelated program that shares the name) is not a usable CLI.
+      const version = /\b(\d+\.\d+\.\d+(?:-[0-9A-Za-z.-]+)?)\b/.exec(output.trim())?.[1] ?? null;
+      finish(version === null
+        ? { ok: false, reason: "no version in --version output" }
+        : { ok: true, version });
+    });
+  });
+}
+
+/**
+ * The system REA this machine already has, or why it was rejected.
+ *
+ * Candidates inside this bundle or this checkout are skipped: a REA that the
+ * provisioner installed is the `bundle` mode, and delegating to it would be a
+ * loop through a path the bundle already covers.
+ */
+async function detectSystemRea({ env = process.env, platform = process.platform, root, pluginRoot = PLUGIN_ROOT } = {}) {
+  const rejected = [];
+  let profiled = 0;
+  const explicit = (env.REA_BUNDLE_SYSTEM_REA ?? "").trim();
+  // Probing costs a process start per candidate, and a row has to answer before
+  // its tools exist. The first existing entry is what the shell would run, so a
+  // couple of fallbacks is plenty; a machine whose first three `rea`s are all
+  // broken is a repair job, not something this decision should keep paying for.
+  const MAX_PROBES = 3;
+  for (const candidate of systemReaCandidates({ env, platform })) {
+    if (isInside(candidate, root) || isInside(candidate, pluginRoot)) {
+      // A REA the provisioner installed is `bundle` mode; delegating to it would
+      // be a loop back into the thing this decision is choosing between.
+      if (rejected.length < 8) rejected.push({ file: candidate, reason: "inside the bundle" });
+      continue;
+    }
+    // The candidate list is PATH x every executable name: one missing name per
+    // directory is normal and is not a finding. Only a regular file that EXISTS
+    // and then fails the probe is worth reporting.
+    try {
+      if (!statSync(candidate).isFile()) continue;
+    } catch {
+      continue;
+    }
+    if (profiled >= MAX_PROBES) {
+      if (rejected.length < 8) rejected.push({ file: candidate, reason: "not probed (budget)" });
+      continue;
+    }
+    profiled += 1;
+    const probe = await probeRea(candidate, { platform });
+    if (probe.ok) return { found: true, file: candidate, version: probe.version };
+    if (rejected.length < 8) rejected.push({ file: candidate, reason: probe.reason });
+  }
+  // Say what actually happened, not what a flag says: an explicit path that was
+  // skipped because it lives inside the bundle is a different answer from one
+  // that does not exist.
+  const explicitPath = explicit === "" ? null : path.resolve(explicit);
+  const explicitRejection = explicitPath === null
+    ? null
+    : (rejected.find(entry => entry.file === explicitPath) ?? null);
+  let reason;
+  if (explicitPath !== null && explicitRejection !== null) {
+    reason = `REA_BUNDLE_SYSTEM_REA was rejected (${explicitRejection.reason}): ${explicitPath}`;
+  } else if (explicitPath !== null && !isFileSync(explicitPath)) {
+    reason = existsSync(explicitPath)
+      ? `REA_BUNDLE_SYSTEM_REA is not a file: ${explicitPath}`
+      : `REA_BUNDLE_SYSTEM_REA does not exist: ${explicitPath}`;
+  } else {
+    reason = profiled === 0 ? "no rea on PATH" : "no usable rea on PATH";
+  }
+  return { found: false, reason, rejected };
+}
+
+/**
+ * Numeric dotted comparison (6.4.0 vs 6.5.0). null when either side is not a
+ * plain x.y.z — a prerelease or a vendor tag is not something this decision
+ * should pretend to order.
+ */
+function compareDotted(a, b) {
+  const parse = (value) => {
+    const match = /^(\d+)\.(\d+)\.(\d+)$/.exec(String(value ?? "").trim());
+    return match ? [Number(match[1]), Number(match[2]), Number(match[3])] : null;
+  };
+  const left = parse(a);
+  const right = parse(b);
+  if (left === null || right === null) return null;
+  for (let index = 0; index < 3; index += 1) {
+    if (left[index] !== right[index]) return left[index] < right[index] ? -1 : 1;
+  }
+  return 0;
+}
+
+/**
+ * Which REA this invocation should run.
+ *
+ * `auto` prefers the system install, but not blindly: a global REA older than
+ * the one this bundle provisioned would silently take the tool count back
+ * (6.4.0 ships 170 tools, 6.5.0 ships 207), so the newer of the two wins and
+ * the older is only used when asked for by name. `bundle` and `system` are
+ * honoured literally 鈥?`system` fails loudly rather than provisioning 800 MB
+ * behind an operator who asked for their own REA.
+ */
+async function resolveLaunch({ root, bundleVersion = null, bundleProvisioned = false, env = process.env, platform = process.platform } = {}) {
+  const preference = launchPreference(env);
+  if (preference.value === "bundle") {
+    return { preference: preference.value, invalid: preference.invalid, mode: "bundle", reason: "pinned by REA_BUNDLE_LAUNCH=bundle" };
+  }
+  // A delegated child that ends up back in this shim (a wrapper script named
+  // `rea` that calls us, a PATH that resolves to a launcher) would recurse
+  // forever: the probe cannot break it, because the grandchild starts a fresh
+  // one. One flag, written into the child's environment, ends the loop — the
+  // nested run takes the bundle path, which either works or fails loudly.
+  if ((env.REA_BUNDLE_DELEGATED ?? "") === "1") {
+    return {
+      preference: preference.value,
+      invalid: preference.invalid,
+      mode: "bundle",
+      reason: "REA_BUNDLE_DELEGATED=1: already running under a delegated system REA, refusing to delegate again",
+    };
+  }
+  const system = await detectSystemRea({ env, platform, root });
+  if (system.found) {
+    const order = bundleVersion === null ? null : compareDotted(system.version, bundleVersion);
+    const pinned = preference.value === "system";
+    // No provisioned build at all: there is nothing to compare against, and the
+    // machine's own REA is exactly what this mode exists to use.
+    if (!bundleProvisioned || pinned || (order !== null && order >= 0)) {
+      return {
+        preference: preference.value,
+        invalid: preference.invalid,
+        mode: "system",
+        reason: pinned
+          ? `pinned by REA_BUNDLE_LAUNCH=system (system ${system.version}, bundle ${bundleVersion ?? "none"})`
+          : !bundleProvisioned
+            ? `a working REA is on PATH (${system.version}) and this machine has no provisioned bundle`
+            : `the REA on PATH (${system.version}) is at least as new as the provisioned one (${bundleVersion})`,
+        system,
+      };
+    }
+    return {
+      preference: preference.value,
+      invalid: preference.invalid,
+      mode: "bundle",
+      reason: order === null
+        ? `the REA on PATH (${system.version}) cannot be compared with the provisioned one (${bundleVersion}); keeping the bundle — set REA_BUNDLE_LAUNCH=system to use the system install`
+        : `the REA on PATH (${system.version}) is older than the provisioned one (${bundleVersion}); set REA_BUNDLE_LAUNCH=system to use it anyway`,
+      system,
+    };
+  }
+  if (preference.value === "system") {
+    return {
+      preference: preference.value,
+      invalid: preference.invalid,
+      mode: "system",
+      reason: `REA_BUNDLE_LAUNCH=system, but no usable REA was found (${system.reason})`,
+      system,
+      unusable: true,
+    };
+  }
+  return {
+    preference: preference.value,
+    invalid: preference.invalid,
+    mode: "bundle",
+    reason: `no usable system REA (${system.reason})`,
+    system,
+  };
+}
+
+/** Launch the system REA exactly the way the 1.x row did. */
+function launchSystemRea(system, argv) {
+  const args = argv.length > 0 ? argv : ["mcp"];
+  const useShell = process.platform === "win32" && /\.(cmd|bat)$/i.test(system.file);
+  // `REA_BUNDLE_DELEGATED` marks the child as "already the system REA". Without
+  // it a wrapper on PATH that calls this shim would delegate to itself forever.
+  const env = { ...process.env, REA_BUNDLE_DELEGATED: "1" };
+  // Only the PATH is quoted. Quoting the arguments too would hand them to a
+  // `.cmd`'s `%*` WITH their quotes (`"mcp"` instead of `mcp`) — measured — which
+  // is a different command line than the 1.x row produced. The row passes `mcp`
+  // and nothing else, so plain tokens are the contract here; an argument that
+  // needs shell quoting is not something this path supports (cmd would also
+  // expand `%NAME%` inside it).
+  const child = useShell
+    ? spawn(comspec(), ["/d", "/s", "/c", `"${[cmdQuote(system.file), ...args].join(" ")}"`], {
+        env,
+        stdio: "inherit",
+        windowsHide: true,
+        windowsVerbatimArguments: true,
+      })
+    : spawn(system.file, args, { env, stdio: "inherit", windowsHide: true });
+  // No signal forwarding on Windows, deliberately. `child.kill()` there is
+  // TerminateProcess on the cmd.exe WRAPPER — the REA inside it survives, still
+  // holding the inherited pipes — and a console Ctrl+C is delivered to the whole
+  // process group anyway, so the child already has it. A hard kill of this shim
+  // (what a supervisor or Task Manager does) runs no JavaScript at all, so the
+  // wrapper must not pretend otherwise: on Windows the delegated REA outlives a
+  // forcibly killed shim. On POSIX the child is in this process's group and gets
+  // the terminal's signal directly; `process.exit` below then ends the shim.
+  child.on("error", (error) => {
+    process.stderr.write(`rea-shim: could not start ${system.file}\n  ${String(error && error.message)}\n`);
+    process.exit(EX_CONFIG);
+  });
+  child.on("exit", (code, signal) => {
+    if (signal) {
+      const signals = { SIGHUP: 1, SIGINT: 2, SIGQUIT: 3, SIGKILL: 9, SIGTERM: 15 };
+      process.exit(128 + (signals[signal] ?? 0));
+    }
+    process.exit(code ?? 0);
+  });
+  return null;
+}
+
+/**
+ * The mirror problem.
+ *
+ * The published row launches `<bundle>/bin/rea-shim.mjs` — a COPY made by
+ * `setup.mjs`. Upgrading the plugin from the market replaces the package but runs
+ * no setup, so the copy keeps executing the code it was made from: every fix in
+ * a new plugin version reaches a user only if they happen to re-run setup. That
+ * is invisible (the old copy keeps working) and it is exactly the wrong
+ * direction for a safety fix.
+ *
+ * The repair has to live in the copy, because the copy is what runs. So the
+ * copy compares itself with the plugin's own `bin/rea-shim.mjs`, recorded in
+ * `mirror.json` by setup, and re-execs that file when the two differ. One hop,
+ * guarded by an environment flag, and only from mirror to checkout — the
+ * checkout copy has no `mirror.json` and would find itself anyway.
+ *
+ * A plugin version older than 2.0.8 cannot do this for its own upgrades: there
+ * is nothing in it to run. `--shim-status` reports `mirror.stale` so the
+ * condition is at least visible, and `check-updates.ps1` says to re-run setup.
+ */
+function sha256(file) {
+  return createHash("sha256").update(readFileSync(file)).digest("hex");
+}
+
+/** Where the installed plugin lives according to setup's record, or null. */
+function pluginCheckout(root) {
+  try {
+    const record = JSON.parse(readFileSync(path.join(root, "mirror.json"), "utf8"));
+    const dir = record?.pluginRoot;
+    if (typeof dir !== "string" || dir.trim() === "") return null;
+    const shim = path.join(dir, "bin", "rea-shim.mjs");
+    return isFileSync(shim) ? shim : null;
+  } catch {
+    return null;
+  }
+}
+
+/**
+ * Whether this copy is behind the plugin it was mirrored from.
+ *
+ * `self` is this file, `source` the checkout's. A checkout copy (it ships
+ * `cordis.patch.yml`) is never stale against itself; a mirror with no record, or
+ * whose recorded checkout is gone, is not re-pointed anywhere.
+ */
+function describeMirror(root, { self = fileURLToPath(import.meta.url) } = {}) {
+  let record = null;
+  try {
+    record = JSON.parse(readFileSync(path.join(root, "mirror.json"), "utf8"));
+  } catch {
+    return null;
+  }
+  const source = pluginCheckout(root);
+  let stale = false;
+  let reason = "no usable plugin checkout in mirror.json";
+  if (source !== null) {
+    if (path.resolve(source) === path.resolve(self)) {
+      reason = "this copy is the plugin checkout";
+    } else {
+      try {
+        stale = sha256(source) !== sha256(self);
+        reason = stale ? "the plugin ships a different launcher than this mirrored copy" : "this copy matches the plugin";
+      } catch {
+        reason = "could not compare the two launchers";
+      }
+    }
+  }
+  return {
+    pluginRoot: typeof record?.pluginRoot === "string" ? record.pluginRoot : null,
+    mirroredAt: typeof record?.mirroredAt === "string" ? record.mirroredAt : null,
+    stale,
+    reason,
+  };
+}
+
+/** Run the checkout's launcher instead of this copy, once. */
+function reexecFromCheckout(source, argv) {
+  const child = spawn(process.execPath, [source, ...argv], {
+    stdio: "inherit",
+    env: { ...process.env, REA_BUNDLE_SHIM_REEXEC: "1" },
+    windowsHide: true,
+  });
+  child.on("error", () => {
+    // Falling back is better than failing: the copy still knows how to launch.
+    process.stderr.write(`rea-shim: could not run ${source}; continuing with the mirrored copy.\n`);
+  });
+  child.on("exit", (code, signal) => {
+    if (signal) {
+      const signals = { SIGHUP: 1, SIGINT: 2, SIGQUIT: 3, SIGKILL: 9, SIGTERM: 15 };
+      process.exit(128 + (signals[signal] ?? 0));
+    }
+    process.exit(code ?? 0);
+  });
+  return null;
+}
+
 /**
  * Usage text, also what a bare invocation prints.
  *
  * `package.json` exposes this file as the `rea-bundle-status` bin, and npm's bin
- * shim can only forward the arguments a user typed — it cannot append a flag. So a
+ * shim can only forward the arguments a user typed 鈥?it cannot append a flag. So a
  * bare `rea-bundle-status`, which is exactly what that name invites, used to fall
  * through to the launch path: it started REA, and on an unprovisioned machine it
  * began an ~800 MB download. Reporting usage instead makes the command's behaviour
@@ -375,19 +943,23 @@ function childEnvironment(config) {
  */
 function usage() {
   return [
-    "rea-shim: launches the vendored REA MCP server for a DSH MCP row.",
+    "rea-shim: launches a REA MCP server for a DSH MCP row.",
     "",
     "  node bin/rea-shim.mjs [rea arguments...]   (normal use: 'mcp')",
-    "  node bin/rea-shim.mjs --shim-status        provisioning status, no launch",
+    "  node bin/rea-shim.mjs --shim-status        launch decision and provisioning status, no launch",
     "  node bin/rea-shim.mjs --shim-print-env     the environment it would export",
     "",
     `  bundle root: ${bundleRoot()}`,
     `  plugin root: ${PLUGIN_ROOT}`,
     "",
+    "  REA_BUNDLE_LAUNCH=auto|system|bundle  (default auto: use the REA already",
+    "  on PATH when it answers --version, otherwise the provisioned bundle)",
+    "  REA_BUNDLE_SYSTEM_REA=<path>          probe this binary instead of PATH",
+    "",
   ].join("\n");
 }
 
-function main() {
+async function main() {
   const argv = process.argv.slice(2);
 
   if (argv.includes("--shim-help") || argv.length === 0) {
@@ -396,11 +968,30 @@ function main() {
   }
 
   const root = bundleRoot();
+  // Self-healing before anything else: a mirrored copy that predates the plugin
+  // it was made from would otherwise keep running that older code forever, and a
+  // market upgrade never re-runs setup. One hop, guarded, and never for the
+  // status query — there the staleness is reported instead.
+  const mirror = describeMirror(root);
+  if (mirror !== null && mirror.stale && process.env.REA_BUNDLE_SHIM_REEXEC !== "1" && !argv.includes("--shim-status")) {
+    const source = pluginCheckout(root);
+    if (source !== null) return reexecFromCheckout(source, argv);
+  }
   const loaded = readBundleConfig(root);
+  // Decide what to run before asking whether the bundle is provisioned: a
+  // machine with its own working REA must not be sent into an ~800 MB
+  // provisioning run just because this bundle has no state yet.
+  const bundleVersion = loaded.ok ? (loaded.config.versions?.reaAgents ?? null) : null;
+  const launch = await resolveLaunch({ root, bundleVersion, bundleProvisioned: loaded.ok });
+  if (launch.invalid !== null) {
+    process.stderr.write(
+      `rea-shim: REA_BUNDLE_LAUNCH="${launch.invalid}" is not one of ${LAUNCH_MODES.join("/")}; using auto.\n`,
+    );
+  }
   // Upstream can add a second row for the same serverName; the shim cannot stop
   // DSH from loading both rows, but saying so is better than doubled tools with
   // no stated cause. The status report carries the same fact for diagnostics.
-  const duplicate = foreignReaRegistration();
+  const duplicate = foreignReaRegistration(root);
   if (duplicate && !argv.includes("--shim-status")) {
     process.stderr.write(
       `\n  dsh-plugin-rea (bundle): duplicate REA registration.\n` +
@@ -417,7 +1008,22 @@ function main() {
       provisioned: loaded.ok,
       reason: loaded.ok ? undefined : loaded.code,
       autoProvision: process.env.REA_BUNDLE_AUTO_PROVISION ?? "(default)",
+      launch: {
+        preference: launch.preference,
+        mode: launch.mode,
+        reason: launch.reason,
+        // Reported, not implied: `mode: "system"` with `unusable: true` is a
+        // request that cannot be met, and a caller reading only the mode would
+        // draw the opposite conclusion.
+        unusable: launch.unusable === true,
+        systemRea: launch.system?.found ? { file: launch.system.file, version: launch.system.version } : null,
+      },
     };
+    if (launch.system?.rejected?.length) report.launch.rejected = launch.system.rejected;
+    // A mirrored copy that lags the plugin keeps working, which is exactly why
+    // nobody notices it: the row runs old code until the mirror is refreshed.
+    // Reported here so `--shim-status` answers "am I running what I installed?".
+    if (mirror !== null) report.mirror = mirror;
     if (duplicate) report.duplicateRegistration = duplicate;
     if (loaded.ok) {
       report.paths = loaded.config.paths ?? {};
@@ -425,7 +1031,33 @@ function main() {
       report.rea = resolveRea(loaded.config, root);
     }
     process.stdout.write(JSON.stringify(report, null, 2) + "\n");
-    return loaded.ok ? 0 : EX_CONFIG;
+    // The row works either way round — a system launch needs no bundle, a bundle
+    // launch needs a provisioned one — but a PINNED system launch with nothing to
+    // launch is a failure whether or not the bundle happens to be provisioned,
+    // and reporting 0 there is how a broken row looks healthy until it starts.
+    const usable = launch.mode === "system" ? launch.unusable !== true : loaded.ok;
+    return usable ? 0 : EX_CONFIG;
+  }
+
+  if (launch.mode === "system") {
+    if (launch.unusable) {
+      process.stderr.write(
+        `\n  dsh-plugin-rea (bundle): REA_BUNDLE_LAUNCH=system was requested, but no usable REA was found.\n` +
+          `  ${launch.reason}\n\n` +
+          `  Install REA (\`npm install -g rea-agents\`), point REA_BUNDLE_SYSTEM_REA at it,\n` +
+          `  or unset REA_BUNDLE_LAUNCH to use the provisioned bundle instead.\n`,
+      );
+      return EX_CONFIG;
+    }
+    if (argv.includes("--shim-print-env")) {
+      process.stdout.write(
+        JSON.stringify({ launch: "system", rea: launch.system.file, version: launch.system.version }, null, 2) + "\n",
+      );
+      return 0;
+    }
+    // The 1.x contract, from the 2.x row: the machine's own REA, its own Node,
+    // its own PATH. Nothing is provisioned and no bundle state is required.
+    return launchSystemRea(launch.system, argv);
   }
 
   if (!loaded.ok) {
@@ -441,12 +1073,12 @@ function main() {
     const provisioning = maybeAutoProvision(root);
     if (provisioning.started) {
       process.stderr.write(
-        `\n  dsh-plugin-rea (bundle): first run — provisioning started in the background.\n  progress: ${provisioning.logPath}\n`,
+        `\n  dsh-plugin-rea (bundle): first run 鈥?provisioning started in the background.\n  progress: ${provisioning.logPath}\n`,
       );
     } else if (provisioning.reason === "last-run-failed") {
       process.stderr.write(
         `\n  dsh-plugin-rea (bundle): the last provisioning attempt failed (exit ${provisioning.exit ?? "?"}).\n` +
-          `  Nothing is running right now — read the log, then re-run setup.\n  log: ${provisioning.logPath}\n`,
+          `  Nothing is running right now 鈥?read the log, then re-run setup.\n  log: ${provisioning.logPath}\n`,
       );
     } else if (provisioning.reason === "in-flight") {
       // A live run owns the lock. Saying nothing here is what made a wedged
@@ -457,7 +1089,7 @@ function main() {
     } else if (provisioning.reason === "spawn-failed" || provisioning.reason === "unwritable") {
       process.stderr.write(
         `\n  dsh-plugin-rea (bundle): provisioning could not be started (${provisioning.reason}).\n` +
-          `  Run it by hand:\n    node "${path.join(root, "scripts", "setup.mjs")}"\n`,
+          `  Run it by hand:\n    node "${path.join(root, "scripts", "setup.mjs")}" --root "${root}"\n`,
       );
     }
     return fail(message, loaded.detail);
@@ -504,5 +1136,8 @@ function main() {
   return null;
 }
 
-const status = main();
+// The launch decision awaits a probe, so `main` is async; the status it returns
+// is this process's exit code, and a launch path resolves to null while the
+// child (already wired to `process.exit`) keeps the loop alive.
+const status = await main();
 if (typeof status === "number") process.exit(status);