@rosthq/cli 0.7.108 → 0.7.109

package/dist/index.js

@@ -50841,7 +50841,7 @@ Every notification should include the seat, cause, evidence, and requested decis
     order: 75,
     title: "Local runner guide",
     summary: "How local agent sessions and runner surfaces should operate through {{brand}} without bypassing Charters or audit.",
-    version: "2026-07-18.1",
+    version: "2026-07-20.1",
     public: true,
     audiences: ["human", "cli", "mcp", "in_app_agent"],
     stages: ["staffing", "operating_rhythm"],
@@ -50887,7 +50887,7 @@ Runner transport defaults to the canonical production host \`https://rost.elevat
 
 Choose the local runtime with \`--runtime auto|claude|codex\` (default \`auto\`). The same short-lived seat-scoped MCP token and server-side guard apply regardless of runtime; results are tagged \`claude-cli\` or \`codex-cli\`.
 
-For always-on execution, \`runner serve --execute\` heartbeats on its normal interval and also checks for claimable work between heartbeats. The default interactive claim cadence is about one second; set \`RUNNER_INTERACTIVE_CLAIM_MS\` or \`--interactive-claim-ms\` to tune it for a host. Claude-backed turns reuse a per-seat Claude Code resume session when Claude reports a session id, so later turns can run with \`claude --resume <session-id>\` under the local subscription account. The runner stores only the opaque resume id in its owner-only state file; it does not use API-billed managed sessions to make subscription runner turns faster. A resident long-lived Claude/Codex process is still prototype-gated, so do not assume a fixed latency SLA from this guide alone.
+For always-on execution, \`runner serve --execute\` heartbeats on its normal interval and also checks for claimable work between heartbeats. The default interactive claim cadence is about one second; set \`RUNNER_INTERACTIVE_CLAIM_MS\` or \`--interactive-claim-ms\` to tune it for a host. Set how many turns the runner may run in parallel with \`--max-sessions <n>\` or \`RUNNER_MAX_SESSIONS\`; left unset, the runner estimates a value from host CPU and memory. The scheduler honors the declared value (clamped to a small ceiling) so a runner never claims more concurrent work than the capacity it reports. Claude-backed turns reuse a per-seat Claude Code resume session when Claude reports a session id, so later turns can run with \`claude --resume <session-id>\` under the local subscription account. The runner stores only the opaque resume id in its owner-only state file; it does not use API-billed managed sessions to make subscription runner turns faster. A resident long-lived Claude/Codex process is still prototype-gated, so do not assume a fixed latency SLA from this guide alone.
 
 The runner bearer secret is stored locally and is never displayed. Each claimed turn receives a short-lived seat-scoped MCP token from {{brand}}, and server-side guards still decide what tools the Seat may use.
 
@@ -57659,7 +57659,7 @@ ${usage()}
   let activeSessions = 0;
   const activeTurnExecutionIds = /* @__PURE__ */ new Set();
   do {
-    const detectedCapabilities = await detectCapabilities(cliVersion, env, homeDir);
+    const detectedCapabilities = await detectCapabilities(cliVersion, env, homeDir, config2.maxSessionsOverride);
     const capabilities = capabilitiesForConfiguredRuntime(detectedCapabilities, config2.runtime);
     const localRuntime = config2.execute ? effectiveExecutionRuntime(config2.runtime, capabilities) : null;
     try {
@@ -57839,6 +57839,7 @@ function parseServeArgs(args, appUrl2, env, homeDir) {
   let once = false;
   let runtime = parseRuntime(env.RUNNER_RUNTIME ?? "auto");
   let userCode = env.RUNNER_USER_CODE;
+  let maxSessionsOverride = parseMaxSessionsEnv(env.RUNNER_MAX_SESSIONS);
   for (let index = 0; index < args.length; index += 1) {
     const token = args[index];
     if (token === "--name") {
@@ -57859,6 +57860,13 @@ function parseServeArgs(args, appUrl2, env, homeDir) {
     } else if (token === "--runtime") {
       runtime = parseRuntime(requiredValue(args, index));
       index += 1;
+    } else if (token === "--max-sessions") {
+      const raw = Number(requiredValue(args, index));
+      if (!Number.isSafeInteger(raw) || raw < 1) {
+        throw new Error("--max-sessions must be an integer >= 1.");
+      }
+      maxSessionsOverride = raw;
+      index += 1;
     } else if (token === "--execute") {
       execute2 = true;
     } else if (token === "--once") {
@@ -57899,9 +57907,17 @@ function parseServeArgs(args, appUrl2, env, homeDir) {
     sandboxAllowNetwork: network.allowNetwork,
     homeDir,
     ...network.unrecognized !== null ? { sandboxNetworkWarning: `RUNNER_SANDBOX_NETWORK='${network.unrecognized}' not recognized (use allow|deny); defaulting to allow egress.` } : {},
-    ...userCode ? { userCode } : {}
+    ...userCode ? { userCode } : {},
+    ...maxSessionsOverride !== void 0 ? { maxSessionsOverride: Math.min(maxSessionsOverride, RUNNER_MAX_SESSIONS_CAP) } : {}
   };
 }
+function parseMaxSessionsEnv(value) {
+  if (value === void 0) {
+    return void 0;
+  }
+  const parsed = Number(value);
+  return Number.isSafeInteger(parsed) && parsed >= 1 ? parsed : void 0;
+}
 function parseRuntime(value) {
   if (value === "auto" || value === "claude" || value === "codex") {
     return value;
@@ -58040,7 +58056,14 @@ async function startPairing(client, config2) {
   }
   return userCode;
 }
-async function detectCapabilities(cliVersion, env, homeDir) {
+var RUNNER_MAX_SESSIONS_CAP = 8;
+function estimateMaxSessions(cpuCount, memoryMb) {
+  const byCpu = Number.isFinite(cpuCount) && cpuCount > 0 ? Math.floor(cpuCount) : 1;
+  const byMem = Number.isFinite(memoryMb) && memoryMb > 0 ? Math.floor(memoryMb / 2048) : 1;
+  const estimate = Math.min(byCpu, byMem);
+  return Math.min(Math.max(estimate, 1), RUNNER_MAX_SESSIONS_CAP);
+}
+async function detectCapabilities(cliVersion, env, homeDir, maxSessionsOverride) {
   const [claude, codex, git, gh, forgeDiskFreeMb] = await Promise.all([
     probeClaude(),
     probe("codex"),
@@ -58049,6 +58072,7 @@ async function detectCapabilities(cliVersion, env, homeDir) {
     probeForgeDiskFreeMb(homeDir)
   ]);
   const accountAlias = resolveAccountAlias(env);
+  const sessionEstimate = estimateMaxSessions(cpus().length, Math.round(totalmem() / (1024 * 1024)));
   return {
     claude: { ...claude, account_alias: accountAlias, rate_limited_until: null },
     codex: { ...codex, account_alias: accountAlias, rate_limited_until: null },
@@ -58061,7 +58085,10 @@ async function detectCapabilities(cliVersion, env, homeDir) {
     ...cliVersion ? { forge_build: cliVersion, seat_work: cliVersion, forge_merge: cliVersion } : {},
     git,
     gh,
-    scheduler: { max_parallel_sessions_estimate: null },
+    scheduler: {
+      max_active_sessions: maxSessionsOverride ?? sessionEstimate,
+      max_parallel_sessions_estimate: sessionEstimate
+    },
     host: { forge_disk_free_mb: forgeDiskFreeMb }
   };
 }
@@ -59408,7 +59435,7 @@ async function clearClaudeSessionId(stateFile, state, workOrder) {
   await saveState(stateFile, state);
 }
 function usage() {
-  return "Usage: rost runner serve [--name <text>] [--state-file <path>] [--heartbeat-ms <n>] [--interactive-claim-ms <n>] [--user-code <code>] [--runtime auto|claude|codex] [--once] [--execute]";
+  return "Usage: rost runner serve [--name <text>] [--state-file <path>] [--heartbeat-ms <n>] [--interactive-claim-ms <n>] [--user-code <code>] [--runtime auto|claude|codex] [--max-sessions <n>] [--once] [--execute]";
 }
 
 // src/index.ts