cool-workflow 0.1.91 → 0.1.93

package/dist/cli/command-surface.js

@@ -43,22 +43,29 @@ const node_child_process_1 = require("node:child_process");
 const readline = __importStar(require("node:readline"));
 const orchestrator_1 = require("../orchestrator");
 const capability_core_1 = require("../capability-core");
-const observability_1 = require("../observability");
-const telemetry_demo_1 = require("../telemetry-demo");
-const run_registry_1 = require("../run-registry");
-const daemon_1 = require("../daemon");
 const scheduler_1 = require("../scheduler");
 const triggers_1 = require("../triggers");
-const workbench_1 = require("../workbench");
-const workbench_host_1 = require("../workbench-host");
+const io_1 = require("./io");
+const run_summary_1 = require("./run-summary");
+const audit_1 = require("./handlers/audit");
+const candidate_1 = require("./handlers/candidate");
+const operator_1 = require("./handlers/operator");
+const registry_1 = require("./handlers/registry");
+const multi_agent_1 = require("./handlers/multi-agent");
+const run_1 = require("./handlers/run");
+const collaboration_1 = require("./handlers/collaboration");
+const blackboard_1 = require("./handlers/blackboard");
+const eval_1 = require("./handlers/eval");
+const node_1 = require("./handlers/node");
+const maintenance_1 = require("./handlers/maintenance");
+const operational_1 = require("./handlers/operational");
+const scheduling_1 = require("./handlers/scheduling");
+const worker_1 = require("./handlers/worker");
+const clones_1 = require("./handlers/clones");
+const workbench_1 = require("./handlers/workbench");
 const operator_ux_1 = require("../operator-ux");
-const multi_agent_operator_ux_1 = require("../multi-agent-operator-ux");
-const multi_agent_eval_1 = require("../multi-agent-eval");
-const state_explosion_1 = require("../state-explosion");
-const evidence_reasoning_1 = require("../evidence-reasoning");
 const doctor_1 = require("../doctor");
 const orchestrator_2 = require("../orchestrator");
-const reporter_1 = require("../reporter");
 const version_1 = require("../version");
 async function runCli(argv = process.argv.slice(2)) {
     const args = (0, orchestrator_1.parseArgv)(argv);
@@ -96,6 +103,13 @@ async function runCli(argv = process.argv.slice(2)) {
         process.env.CW_NO_COLOR = "1";
     if (args.options.full)
         process.env.CW_OUTPUT = "full";
+    // `cw <verb> --help` / `-h` -> per-command help (the verb's subcommands +
+    // one-line summaries), derived from the capability registry. Additive: the
+    // bare `cw` / `cw --help` top-level help is handled above.
+    if ((args.options.help || args.options.h) && args.command && !args.command.startsWith("-")) {
+        process.stdout.write((0, orchestrator_1.formatCommandHelp)(args.command) + "\n");
+        return;
+    }
     // Bare -q / --question -> redirect to quickstart (auto-detect repo/agent/app).
     // CONSUME the positional (shift) so the question never survives as positionals[0]
     // — otherwise the quickstart handler reads it as the app id ("Workflow app not found").
@@ -113,7 +127,11 @@ async function runCli(argv = process.argv.slice(2)) {
     const scheduler = new scheduler_1.Scheduler(String(args.options.cwd || process.cwd()));
     const triggers = new triggers_1.RoutineTriggerBridge(String(args.options.cwd || process.cwd()));
     switch (args.command) {
-        case "help":
+        case "help": {
+            const [topic] = args.positionals;
+            process.stdout.write((topic ? (0, orchestrator_1.formatCommandHelp)(topic) : (0, orchestrator_1.formatHelp)()) + "\n");
+            return;
+        }
         case undefined:
             process.stdout.write((0, orchestrator_1.formatHelp)() + "\n");
             return;
@@ -141,7 +159,7 @@ async function runCli(argv = process.argv.slice(2)) {
             return;
         }
         case "list":
-            printJson(runner.listWorkflows());
+            (0, io_1.printJson)(runner.listWorkflows());
             return;
         case "search": {
             const keyword = args.positionals.join(" ");
@@ -150,8 +168,8 @@ async function runCli(argv = process.argv.slice(2)) {
             const apps = runner.listApps();
             const lower = keyword.toLowerCase();
             const results = apps.filter((a) => a.title.toLowerCase().includes(lower) || a.summary.toLowerCase().includes(lower) || a.id.toLowerCase().includes(lower)).map((a) => ({ id: a.id, title: a.title, summary: a.summary }));
-            if (wantsJson(args.options))
-                printJson(results);
+            if ((0, io_1.wantsJson)(args.options))
+                (0, io_1.printJson)(results);
             else
                 process.stdout.write(`${(0, orchestrator_2.formatSearchResults)(keyword, results)}\n`);
             return;
@@ -186,16 +204,16 @@ async function runCli(argv = process.argv.slice(2)) {
             if (!appId)
                 throw new Error("Missing workflow app id.\n  Tip: list apps with \"cw list\", then \"cw info <id>\" for details");
             const data = runner.showApp(appId);
-            if (wantsJson(args.options))
-                printJson(data);
+            if ((0, io_1.wantsJson)(args.options))
+                (0, io_1.printJson)(data);
             else
                 process.stdout.write(`${(0, orchestrator_2.formatInfo)(appId, data)}\n`);
             return;
         }
         case "doctor": {
             const report = (0, doctor_1.runDoctor)(args.options, process.env, String(args.options.cwd || process.cwd()));
-            if (wantsJson(args.options))
-                printJson(report);
+            if ((0, io_1.wantsJson)(args.options))
+                (0, io_1.printJson)(report);
             else if (args.options.fix)
                 process.stdout.write(`${(0, doctor_1.formatDoctorFixes)(report)}\n`);
             else
@@ -208,33 +226,33 @@ async function runCli(argv = process.argv.slice(2)) {
             const [workflowId] = args.positionals;
             if (!workflowId)
                 throw new Error("Missing workflow id.\n  Tip: create one with \"cw init my-workflow\" or list with \"cw list\"");
-            printJson(runner.init(workflowId, args.options));
+            (0, io_1.printJson)(runner.init(workflowId, args.options));
             return;
         }
         case "app": {
             const [subcommand, appIdOrPath] = args.positionals;
             switch (subcommand) {
                 case "list":
-                    printJson(runner.listApps());
+                    (0, io_1.printJson)(runner.listApps());
                     return;
                 case "show":
-                    printJson(runner.showApp(required(appIdOrPath, "app id")));
+                    (0, io_1.printJson)(runner.showApp((0, io_1.required)(appIdOrPath, "app id")));
                     return;
                 case "validate": {
-                    const result = runner.validateApp(required(appIdOrPath, "app path or id"));
-                    printJson(result);
+                    const result = runner.validateApp((0, io_1.required)(appIdOrPath, "app path or id"));
+                    (0, io_1.printJson)(result);
                     if (!result.valid)
                         process.exitCode = 1;
                     return;
                 }
                 case "init":
-                    printJson(runner.initApp(required(appIdOrPath, "app id"), args.options));
+                    (0, io_1.printJson)(runner.initApp((0, io_1.required)(appIdOrPath, "app id"), args.options));
                     return;
                 case "package":
-                    printJson(runner.packageApp(required(appIdOrPath, "app id"), args.options));
+                    (0, io_1.printJson)(runner.packageApp((0, io_1.required)(appIdOrPath, "app id"), args.options));
                     return;
                 case "run":
-                    printJson((0, capability_core_1.appRun)(runner, { ...args.options, appId: required(appIdOrPath, "app id") }));
+                    (0, io_1.printJson)((0, capability_core_1.appRun)(runner, { ...args.options, appId: (0, io_1.required)(appIdOrPath, "app id") }));
                     return;
                 default:
                     throw new Error("Usage: cw.js app list|show|validate|init|package|run [app-id|path]");
@@ -247,18 +265,18 @@ async function runCli(argv = process.argv.slice(2)) {
             // DELEGATES worker execution to the operator's configured agent backend and
             // fails closed (status=blocked) when none is set. No new executor/scheduler.
             const [appId] = args.positionals;
-            const runId = optionalArg(args.options.run) || optionalArg(args.options.runId);
+            const runId = (0, io_1.optionalArg)(args.options.run) || (0, io_1.optionalArg)(args.options.runId);
             await promptQuestion(args.options);
             const qs = (0, capability_core_1.quickstart)(runner, { ...args.options, ...(appId ? { appId } : {}), ...(runId ? { runId } : {}) });
-            printJson(qs);
+            (0, io_1.printJson)(qs);
             const qr = qs;
             // Clean human summary on stderr (TTY-gated, inside the reporter). Suppressed under --json so
             // machine mode emits ONLY the stdout payload — no stderr chrome to parse around. The type
             // guard also skips --check/--preview results (no reportPath of their own). The summary is the
             // COMPACT findings table (re-parsed from each completed worker's cw:result), the report path,
             // and where the per-worker transcripts live — NOT the full prose (that's report.md/--full).
-            if (!wantsJson(args.options) && typeof qr.runId === "string" && typeof qr.reportPath === "string") {
-                emitRunSummary(runner, args.options, {
+            if (!(0, io_1.wantsJson)(args.options) && typeof qr.runId === "string" && typeof qr.reportPath === "string") {
+                (0, run_summary_1.emitRunSummary)(runner, args.options, {
                     runId: qr.runId,
                     reportPath: qr.reportPath,
                     status: String(qr.status || ""),
@@ -283,41 +301,41 @@ async function runCli(argv = process.argv.slice(2)) {
             const [workflowId] = args.positionals;
             if (!workflowId)
                 throw new Error("Missing workflow id.\n  Tip: plan an architecture review with \"cw plan architecture-review\"");
-            printJson((0, capability_core_1.planSummary)(runner, workflowId, args.options));
+            (0, io_1.printJson)((0, capability_core_1.planSummary)(runner, workflowId, args.options));
             return;
         }
         case "status":
             if (!args.positionals[0]) {
                 const nextActions = (0, operator_ux_1.adviseNoRun)();
-                if (wantsJson(args.options))
-                    printJson({ runId: null, nextActions });
+                if ((0, io_1.wantsJson)(args.options))
+                    (0, io_1.printJson)({ runId: null, nextActions });
                 else
                     process.stdout.write(`No run selected\n\nNext Action\n${nextActions.map((action) => `  ${action.command}\n    reason: ${action.reason}`).join("\n")}\n`);
             }
-            else if (wantsJson(args.options))
-                printJson(runner.status(args.positionals[0]));
+            else if ((0, io_1.wantsJson)(args.options))
+                (0, io_1.printJson)(runner.status(args.positionals[0]));
             else {
                 const summary = runner.operatorStatus(args.positionals[0]);
                 process.stdout.write(`${(args.options.summary || args.options.brief ? (0, operator_ux_1.formatOperatorSummary)(summary) : (0, operator_ux_1.formatOperatorStatus)(summary))}\n`);
             }
             return;
         case "next":
-            printJson(runner.next(required(args.positionals[0], "run id"), args.options));
+            (0, io_1.printJson)(runner.next((0, io_1.required)(args.positionals[0], "run id"), args.options));
             return;
         case "dispatch":
-            printJson(runner.dispatch(required(args.positionals[0], "run id"), args.options));
+            (0, io_1.printJson)(runner.dispatch((0, io_1.required)(args.positionals[0], "run id"), args.options));
             return;
         case "result": {
             const [runId, taskId, resultPath] = args.positionals;
-            printJson(runner.recordResult(required(runId, "run id"), required(taskId, "task id"), required(resultPath, "result file"), args.options));
+            (0, io_1.printJson)(runner.recordResult((0, io_1.required)(runId, "run id"), (0, io_1.required)(taskId, "task id"), (0, io_1.required)(resultPath, "result file"), args.options));
             return;
         }
         case "state": {
             const [subcommand, runId] = args.positionals;
             switch (subcommand) {
                 case "check": {
-                    const report = runner.checkState(required(runId, "run id"), args.options);
-                    printJson(report);
+                    const report = runner.checkState((0, io_1.required)(runId, "run id"), args.options);
+                    (0, io_1.printJson)(report);
                     if (report.status === "unsupported")
                         process.exitCode = 1;
                     return;
@@ -328,1323 +346,132 @@ async function runCli(argv = process.argv.slice(2)) {
         }
         case "commit":
             if (args.positionals[0] === "summary") {
-                const summary = runner.summarizeCommitRecords(required(args.positionals[1], "run id"));
-                if (wantsJson(args.options))
-                    printJson(summary);
+                const summary = runner.summarizeCommitRecords((0, io_1.required)(args.positionals[1], "run id"));
+                if ((0, io_1.wantsJson)(args.options))
+                    (0, io_1.printJson)(summary);
                 else
                     process.stdout.write(`${(0, operator_ux_1.formatCommitSummary)(summary)}\n`);
                 return;
             }
-            printJson(runner.commit(required(args.positionals[0], "run id"), args.options));
+            (0, io_1.printJson)(runner.commit((0, io_1.required)(args.positionals[0], "run id"), args.options));
             return;
-        case "report": {
-            // `report verify-bundle <path>` is the offline self-contained bundle verifier;
-            // `report bundle <run-id>` exports a sealed bundle and self-verifies it;
-            // every other `report <run-id>` form prints/inspects a local run's report.
-            if (args.positionals[0] === "verify-bundle") {
-                const result = (0, capability_core_1.runVerifyReportBundle)(runner, { ...args.options, archive: args.positionals[1] || args.options.archive || args.options.path || args.options.file || args.options.bundle });
-                printJson(result);
-                // Fail closed: a forged/edited/corrupt bundle verifies false — surface it
-                // through the exit code so `cw report verify-bundle <file> && ship` cannot
-                // pass on a lie. Mirrors run inspect-archive / telemetry verify.
-                if (!result.ok)
-                    process.exitCode = 1;
-                return;
-            }
-            if (args.positionals[0] === "bundle") {
-                const result = (0, capability_core_1.reportBundle)(runner, required(args.positionals[1] || optionalArg(args.options.runId || args.options.run), "run id"), args.options);
-                printJson(result);
-                // Fail closed: never report a "bundle made" success if the artifact does not
-                // self-verify — so `cw report bundle <run> && send-to-client` cannot ship an
-                // unverifiable report (e.g. no trust key under --strict-signatures).
-                if (!result.ok)
-                    process.exitCode = 1;
-                return;
-            }
-            const runId = required(args.positionals[0], "run id");
-            const report = runner.report(runId);
-            if (wantsJson(args.options)) {
-                printJson(report);
-            }
-            else if (args.options.show || args.options.summary) {
-                process.stdout.write(`${(0, operator_ux_1.formatOperatorReport)(runner.operatorReport(runId))}\n`);
-                process.stdout.write(`\n${(0, state_explosion_1.formatStateExplosionReport)(runner.stateExplosionReport(runId))}\n`);
-            }
-            else {
-                process.stdout.write(`${report.path}\n`);
-            }
+        case "report":
+            (0, operator_1.handleReport)(args, runner);
             return;
-        }
-        case "operator": {
-            const [subcommand, runId] = args.positionals;
-            switch (subcommand) {
-                case "status":
-                    if (wantsJson(args.options))
-                        printJson(runner.operatorStatus(required(runId, "run id")));
-                    else {
-                        const summary = runner.operatorStatus(required(runId, "run id"));
-                        process.stdout.write(`${(args.options.summary || args.options.brief ? (0, operator_ux_1.formatOperatorSummary)(summary) : (0, operator_ux_1.formatOperatorStatus)(summary))}\n`);
-                    }
-                    return;
-                case "report":
-                    if (wantsJson(args.options))
-                        printJson(runner.operatorReport(required(runId, "run id")));
-                    else
-                        process.stdout.write(`${(0, operator_ux_1.formatOperatorReport)(runner.operatorReport(required(runId, "run id")))}\n`);
-                    return;
-                default:
-                    throw new Error("Usage: cw.js operator status|report <run-id> [--json]");
-            }
-        }
-        case "graph": {
-            const graph = runner.operatorGraph(required(args.positionals[0], "run id"));
-            if (wantsJson(args.options))
-                printJson(graph);
-            else
-                process.stdout.write(`${(0, operator_ux_1.formatOperatorGraph)(graph)}\n`);
+        case "operator":
+            (0, operator_1.handleOperator)(args, runner);
             return;
-        }
-        case "topology": {
-            const [subcommand, first, second] = args.positionals;
-            switch (subcommand) {
-                case "list":
-                    printJson(runner.listTopologies());
-                    return;
-                case "show":
-                    if (second)
-                        printJson(runner.showTopologyRun(required(first, "run id"), second));
-                    else
-                        printJson(runner.showTopology(required(first, "topology id")));
-                    return;
-                case "validate": {
-                    const result = runner.validateTopology(required(first, "topology id"));
-                    printJson(result);
-                    if (!result.valid)
-                        process.exitCode = 1;
-                    return;
-                }
-                case "apply":
-                    printJson(runner.applyTopology(required(first, "run id"), required(second, "topology id"), args.options));
-                    return;
-                case "summary": {
-                    const summary = runner.topologySummary(required(first, "run id"));
-                    if (wantsJson(args.options))
-                        printJson(summary);
-                    else
-                        process.stdout.write(`${(0, operator_ux_1.formatTopologySummary)(summary)}\n`);
-                    return;
-                }
-                case "graph": {
-                    const graph = runner.topologyGraph(required(first, "run id"));
-                    if (wantsJson(args.options))
-                        printJson(graph);
-                    else
-                        process.stdout.write(`${(0, operator_ux_1.formatOperatorGraph)({ runId: required(first, "run id"), nodes: graph.nodes, edges: graph.edges })}\n`);
-                    return;
-                }
-                default:
-                    throw new Error("Usage: cw.js topology list|show <topology-id>|show <run-id> <topology-run-id>|validate <topology-id>|apply <run-id> <topology-id>|summary <run-id>|graph <run-id>");
-            }
-        }
-        case "summary": {
-            const [subcommand, runId] = args.positionals;
-            switch (subcommand) {
-                case "refresh": {
-                    const index = runner.summaryRefresh(required(runId, "run id"), args.options);
-                    if (wantsJson(args.options))
-                        printJson(index);
-                    else
-                        process.stdout.write(`${(0, state_explosion_1.formatStateExplosionReport)(runner.summaryShow(required(runId, "run id")))}\n`);
-                    return;
-                }
-                case "show": {
-                    const report = runner.summaryShow(required(runId, "run id"));
-                    if (wantsJson(args.options))
-                        printJson(report);
-                    else
-                        process.stdout.write(`${(0, state_explosion_1.formatStateExplosionReport)(report)}\n`);
-                    return;
-                }
-                default:
-                    throw new Error("Usage: cw.js summary refresh|show <run-id> [--json]");
-            }
-        }
-        case "multi-agent": {
-            const [subcommand, runId, id] = args.positionals;
-            switch (subcommand) {
-                case "status":
-                    if (wantsJson(args.options))
-                        printJson(runner.hostMultiAgentStatus(required(runId, "run id")));
-                    else
-                        process.stdout.write(`${(0, multi_agent_operator_ux_1.formatMultiAgentOperatorStatus)(runner.multiAgentOperatorStatus(required(runId, "run id")))}\n`);
-                    return;
-                case "step":
-                    printJson(runner.hostMultiAgentStep(required(runId, "run id"), args.options));
-                    return;
-                case "blackboard":
-                    printJson(runner.hostMultiAgentBlackboard(required(runId, "run id"), id, args.options));
-                    return;
-                case "score":
-                    printJson(runner.hostMultiAgentScore(required(runId, "run id"), { ...args.options, candidate: args.options.candidate || args.options.candidateId || id }));
-                    return;
-                case "select":
-                    printJson(runner.hostMultiAgentSelect(required(runId, "run id"), { ...args.options, candidate: args.options.candidate || args.options.candidateId || id }));
-                    return;
-                case "summary": {
-                    const summary = runner.multiAgentSummary(required(runId, "run id"));
-                    if (wantsJson(args.options))
-                        printJson(summary);
-                    else
-                        process.stdout.write(`${(0, operator_ux_1.formatMultiAgentSummary)(summary)}\n`);
-                    return;
-                }
-                case "summarize": {
-                    const report = runner.multiAgentSummarize(required(runId, "run id"));
-                    if (wantsJson(args.options))
-                        printJson(report);
-                    else
-                        process.stdout.write(`${(0, state_explosion_1.formatStateExplosionReport)(report)}\n`);
-                    return;
-                }
-                case "graph": {
-                    const wantsView = args.options.view || args.options.focus || args.options.depth;
-                    if (wantsView) {
-                        const graph = runner.multiAgentGraphView(required(runId, "run id"), args.options);
-                        if (wantsJson(args.options))
-                            printJson(graph);
-                        else
-                            process.stdout.write(`${(0, state_explosion_1.formatCompactGraph)(graph)}\n`);
-                        return;
-                    }
-                    const graph = runner.multiAgentOperatorGraph(required(runId, "run id"));
-                    if (wantsJson(args.options))
-                        printJson(graph);
-                    else
-                        process.stdout.write(`${(0, operator_ux_1.formatOperatorGraph)({ runId: required(runId, "run id"), nodes: graph.nodes, edges: graph.edges })}\n`);
-                    return;
-                }
-                case "dependencies": {
-                    const rows = runner.multiAgentDependencies(required(runId, "run id"));
-                    if (wantsJson(args.options))
-                        printJson(rows);
-                    else
-                        process.stdout.write(`${(0, multi_agent_operator_ux_1.formatMultiAgentDependencies)(rows)}\n`);
-                    return;
-                }
-                case "failures": {
-                    const rows = runner.multiAgentFailures(required(runId, "run id"));
-                    if (wantsJson(args.options))
-                        printJson(rows);
-                    else
-                        process.stdout.write(`${(0, multi_agent_operator_ux_1.formatMultiAgentFailures)(rows)}\n`);
-                    return;
-                }
-                case "evidence": {
-                    const rows = runner.multiAgentEvidence(required(runId, "run id"));
-                    if (wantsJson(args.options))
-                        printJson(rows);
-                    else
-                        process.stdout.write(`${(0, multi_agent_operator_ux_1.formatMultiAgentEvidence)(rows)}\n`);
-                    return;
-                }
-                case "reasoning": {
-                    if (args.options.refresh && !args.options.evidence && !args.options.evidenceId) {
-                        const index = runner.multiAgentReasoningRefresh(required(runId, "run id"));
-                        printJson(index);
-                        return;
-                    }
-                    const report = runner.multiAgentReasoning(required(runId, "run id"), { ...args.options, evidence: args.options.evidence || args.options.evidenceId || id });
-                    if (wantsJson(args.options))
-                        printJson(report);
-                    else
-                        process.stdout.write(`${(0, evidence_reasoning_1.formatEvidenceReasoningReport)(report)}\n`);
-                    return;
-                }
-                case "run":
-                    if (!runId ||
-                        args.options.topology ||
-                        args.options.topologyId ||
-                        args.options.app ||
-                        args.options.appId ||
-                        args.options.workflow ||
-                        args.options.workflowId) {
-                        printJson(runner.hostMultiAgentRun(runId, args.options));
-                        return;
-                    }
-                    if (id && !args.options.id && !args.options.status)
-                        printJson(runner.showMultiAgentRun(required(runId, "run id"), id));
-                    else if (id && args.options.status)
-                        printJson(runner.transitionMultiAgentRun(required(runId, "run id"), id, args.options));
-                    else
-                        printJson(runner.createMultiAgentRun(required(runId, "run id"), args.options));
-                    return;
-                case "show":
-                    printJson(runner.showMultiAgentRun(required(runId, "run id"), required(id, "multi-agent run id")));
-                    return;
-                case "role":
-                    if (id && !args.options.id && !args.options["multi-agent-run"] && !args.options.multiAgentRun && !args.options.multiAgentRunId) {
-                        printJson(runner.showAgentRole(required(runId, "run id"), id));
-                    }
-                    else {
-                        printJson(runner.createAgentRole(required(runId, "run id"), { ...args.options, id: args.options.id || id }));
-                    }
-                    return;
-                case "group":
-                    if (id && !args.options.id && !args.options["multi-agent-run"] && !args.options.multiAgentRun && !args.options.multiAgentRunId) {
-                        printJson(runner.showAgentGroup(required(runId, "run id"), id));
-                    }
-                    else {
-                        printJson(runner.createAgentGroup(required(runId, "run id"), { ...args.options, id: args.options.id || id }));
-                    }
-                    return;
-                case "membership":
-                    if (id && !args.options.id && !args.options.group && !args.options.groupId && !args.options["multi-agent-group"]) {
-                        printJson(runner.showAgentMembership(required(runId, "run id"), id));
-                    }
-                    else {
-                        printJson(runner.assignAgentMembership(required(runId, "run id"), { ...args.options, id: args.options.id || id }));
-                    }
-                    return;
-                case "fanout":
-                    if (id && !args.options.id && !args.options.group && !args.options.groupId && !args.options["multi-agent-group"]) {
-                        printJson(runner.showAgentFanout(required(runId, "run id"), id));
-                    }
-                    else {
-                        printJson(runner.createAgentFanout(required(runId, "run id"), { ...args.options, id: args.options.id || id }));
-                    }
-                    return;
-                case "fanin":
-                    if (id && !args.options.id && !args.options.group && !args.options.groupId && !args.options["multi-agent-group"] && !args.options.fanout) {
-                        printJson(runner.showAgentFanin(required(runId, "run id"), id));
-                    }
-                    else {
-                        printJson(runner.collectAgentFanin(required(runId, "run id"), { ...args.options, id: args.options.id || id }));
-                    }
-                    return;
-                default:
-                    throw new Error("Usage: cw.js multi-agent run|status|step|blackboard|score|select|summary|summarize|graph|dependencies|failures|evidence|reasoning|show|role|group|membership|fanout|fanin <run-id> [id]");
-            }
-        }
-        case "eval": {
-            const [subcommand, first, second] = args.positionals;
-            let result;
-            switch (subcommand) {
-                case "snapshot":
-                    result = runner.evalSnapshot(required(first, "run id"), args.options);
-                    break;
-                case "replay":
-                    result = runner.evalReplay(required(first, "snapshot id or path"), args.options);
-                    break;
-                case "compare":
-                    result = runner.evalCompare(required(first, "baseline id or path"), required(second, "replay id or path"));
-                    break;
-                case "score":
-                    result = runner.evalScore(required(first, "replay id or path"));
-                    break;
-                case "gate":
-                    result = runner.evalGate(required(first, "suite id or path"));
-                    if (!wantsJson(args.options) && result.status === "fail")
-                        process.exitCode = 1;
-                    break;
-                case "report":
-                    result = runner.evalReport(required(first, "replay id or path"));
-                    break;
-                default:
-                    throw new Error("Usage: cw.js eval snapshot <run-id> --id <snapshot-id> | replay <snapshot-id-or-path> | compare <baseline-id-or-path> <replay-id-or-path> | score <replay-id-or-path> | gate <suite-id-or-path> | report <replay-id-or-path>");
-            }
-            if (wantsJson(args.options))
-                printJson(result);
-            else
-                process.stdout.write(`${(0, multi_agent_eval_1.formatMultiAgentEval)(result)}\n`);
-            if (subcommand === "gate" && result.status === "fail")
-                process.exitCode = 1;
+        case "graph":
+            (0, operator_1.handleGraph)(args, runner);
+            return;
+        case "topology":
+            (0, operator_1.handleTopology)(args, runner);
+            return;
+        case "summary":
+            (0, operator_1.handleSummary)(args, runner);
+            return;
+        case "multi-agent":
+            (0, multi_agent_1.handleMultiAgent)(args, runner);
+            return;
+        case "eval":
+            (0, eval_1.handleEval)(args, runner);
+            return;
+        case "blackboard":
+            (0, blackboard_1.handleBlackboard)(args, runner);
+            return;
+        case "coordinator":
+            (0, blackboard_1.handleCoordinator)(args, runner);
+            return;
+        case "sandbox":
+            (0, operational_1.handleSandbox)(args, runner);
+            return;
+        case "backend":
+            (0, operational_1.handleBackend)(args, runner);
+            return;
+        case "contract":
+            (0, operational_1.handleContract)(args, runner);
+            return;
+        case "node":
+            (0, node_1.handleNode)(args, runner);
+            return;
+        case "migration":
+            (0, operational_1.handleMigration)(args, runner);
+            return;
+        case "feedback":
+            (0, operational_1.handleFeedback)(args, runner);
+            return;
+        case "worker":
+            (0, worker_1.handleWorker)(args, runner);
+            return;
+        case "audit":
+            (0, audit_1.handleAudit)(args, runner);
+            return;
+        case "candidate":
+            (0, candidate_1.handleCandidate)(args, runner);
             return;
-        }
-        case "blackboard": {
-            const [subcommand, action, runId] = args.positionals;
-            switch (subcommand) {
-                case "summary":
-                    printJson(runner.blackboardSummary(required(action, "run id"), args.options));
-                    return;
-                case "summarize": {
-                    const digest = runner.blackboardSummarize(required(action, "run id"), args.options);
-                    if (wantsJson(args.options))
-                        printJson(digest);
-                    else
-                        process.stdout.write(`${(0, state_explosion_1.formatBlackboardDigest)(digest)}\n`);
-                    return;
-                }
-                case "graph":
-                    printJson(runner.blackboardGraph(required(action, "run id")));
-                    return;
-                case "resolve":
-                    printJson(runner.resolveRunBlackboard(required(action, "run id"), args.options));
-                    return;
-                case "topic":
-                    if (action === "create") {
-                        printJson(runner.createBlackboardTopic(required(runId, "run id"), args.options));
-                        return;
-                    }
-                    break;
-                case "message":
-                    if (action === "post") {
-                        printJson(runner.postBlackboardMessage(required(runId, "run id"), args.options));
-                        return;
-                    }
-                    if (action === "list") {
-                        printJson(runner.listBlackboardMessages(required(runId, "run id"), args.options));
-                        return;
-                    }
-                    break;
-                case "context":
-                    if (action === "put") {
-                        printJson(runner.putBlackboardContext(required(runId, "run id"), args.options));
-                        return;
-                    }
-                    break;
-                case "artifact":
-                    if (action === "add") {
-                        printJson(runner.addBlackboardArtifact(required(runId, "run id"), args.options));
-                        return;
-                    }
-                    if (action === "list") {
-                        printJson(runner.listBlackboardArtifacts(required(runId, "run id"), args.options));
-                        return;
-                    }
-                    break;
-                case "snapshot":
-                    printJson(runner.snapshotBlackboard(required(action, "run id"), args.options));
-                    return;
-                default:
-                    break;
-            }
-            throw new Error("Usage: cw.js blackboard summary|summarize|graph|resolve <run-id> | topic create <run-id> | message post|list <run-id> | context put <run-id> | artifact add|list <run-id> | snapshot <run-id>");
-        }
-        case "coordinator": {
-            const [subcommand, runId] = args.positionals;
-            switch (subcommand) {
-                case "summary":
-                    printJson(runner.coordinatorSummary(required(runId, "run id"), args.options));
-                    return;
-                case "decision":
-                    printJson(runner.recordCoordinatorDecision(required(runId, "run id"), args.options));
-                    return;
-                default:
-                    throw new Error("Usage: cw.js coordinator summary <run-id> | coordinator decision <run-id> --kind <kind> --outcome <outcome> --reason TEXT");
-            }
-        }
-        case "sandbox": {
-            const [subcommand, profileIdOrFile] = args.positionals;
-            switch (subcommand) {
-                case "list":
-                    printJson(runner.listSandboxProfiles(args.options));
-                    return;
-                case "show":
-                    printJson(runner.showSandboxProfile(required(profileIdOrFile, "profile id"), args.options));
-                    return;
-                case "validate": {
-                    const result = runner.validateSandboxProfile(required(profileIdOrFile, "profile file"), args.options);
-                    printJson(result);
-                    if (!result.valid)
-                        process.exitCode = 1;
-                    return;
-                }
-                case "choose":
-                case "resolve":
-                    printJson((0, capability_core_1.sandboxChoose)(runner, { ...args.options, profileId: profileIdOrFile || args.options.profileId }));
-                    return;
-                default:
-                    throw new Error("Usage: cw.js sandbox list|show|validate|choose|resolve [profile-id|profile-file]");
-            }
-        }
-        case "backend": {
-            const [subcommand, backendId] = args.positionals;
-            switch (subcommand) {
-                case "list":
-                    printJson(runner.listBackends(args.options));
-                    return;
-                case "show":
-                    printJson(runner.showBackend(required(backendId, "backend id"), args.options));
-                    return;
-                case "probe":
-                    printJson(runner.probeBackend(backendId, args.options));
-                    return;
-                case "agent": {
-                    // `backend agent config [show]` = read-only; `backend agent config set ...` = mutating.
-                    const [, , action] = args.positionals;
-                    if (action === "set") {
-                        printJson((0, capability_core_1.backendAgentConfigSet)(args.options));
-                        return;
-                    }
-                    printJson((0, capability_core_1.backendAgentConfigShow)(args.options));
-                    return;
-                }
-                default:
-                    throw new Error("Usage: cw.js backend list|show|probe [backend-id]  |  cw.js backend agent config [show|set] [--agent-command ... --agent-endpoint ... --agent-model ...]");
-            }
-        }
-        case "contract": {
-            const [subcommand, runId, contractId] = args.positionals;
-            switch (subcommand) {
-                case "show":
-                    printJson(runner.showContract(required(runId, "run id"), contractId));
-                    return;
-                default:
-                    throw new Error("Usage: cw.js contract show <run-id> [contract-id]");
-            }
-        }
-        case "node": {
-            const [subcommand, runId, nodeId] = args.positionals;
-            switch (subcommand) {
-                case "list":
-                    printJson(runner.listNodes(required(runId, "run id")));
-                    return;
-                case "show":
-                    printJson(runner.showNode(required(runId, "run id"), required(nodeId, "node id")));
-                    return;
-                case "graph":
-                    if (wantsJson(args.options))
-                        printJson(runner.graphNodes(required(runId, "run id")));
-                    else
-                        process.stdout.write(`${(0, operator_ux_1.formatOperatorGraph)(runner.operatorGraph(required(runId, "run id")))}\n`);
-                    return;
-                case "snapshot":
-                    printJson(runner.nodeSnapshot(required(runId, "run id"), required(nodeId, "node id")));
-                    return;
-                case "diff":
-                    printJson(runner.nodeDiff(required(runId, "run id"), required(nodeId, "baseline snapshot id"), required(args.positionals[3], "candidate snapshot id")));
-                    return;
-                case "replay":
-                    printJson(runner.nodeReplay(required(runId, "run id"), required(nodeId, "snapshot id")));
-                    return;
-                case "verify": {
-                    const verdict = runner.nodeReplayVerify(required(runId, "run id"), required(nodeId, "replay id"));
-                    printJson(verdict);
-                    if (!verdict.pass)
-                        process.exitCode = 1;
-                    return;
-                }
-                default:
-                    throw new Error("Usage: cw.js node list|show|graph|snapshot|diff|replay|verify <run-id> [node-id|snapshot-id|replay-id]");
-            }
-        }
-        case "migration": {
-            const [subcommand, target] = args.positionals;
-            switch (subcommand) {
-                case "list":
-                    printJson(runner.migrationList());
-                    return;
-                case "check": {
-                    const report = runner.migrationCheck(required(target, "target (run-id or state/app file)"), args.options);
-                    printJson(report);
-                    if (report.status === "unsupported")
-                        process.exitCode = 1;
-                    return;
-                }
-                case "prove": {
-                    const proof = runner.migrationProve(required(target, "target (run-id or state/app file)"), args.options);
-                    printJson(proof);
-                    if (!proof.pass)
-                        process.exitCode = 1;
-                    return;
-                }
-                default:
-                    throw new Error("Usage: cw.js migration list|check|prove [target] [--contract run-state|workflow-app]");
-            }
-        }
-        case "feedback": {
-            const [subcommand, runId, feedbackId] = args.positionals;
-            switch (subcommand) {
-                case "list":
-                    printJson(runner.listFeedback(required(runId, "run id"), args.options));
-                    return;
-                case "show":
-                    printJson(runner.showFeedback(required(runId, "run id"), required(feedbackId, "feedback id")));
-                    return;
-                case "collect":
-                    printJson(runner.collectFeedback(required(runId, "run id")));
-                    return;
-                case "summary": {
-                    const summary = runner.summarizeFeedbackRecords(required(runId, "run id"));
-                    if (wantsJson(args.options))
-                        printJson(summary);
-                    else
-                        process.stdout.write(`${(0, operator_ux_1.formatFeedbackSummary)(summary)}\n`);
-                    return;
-                }
-                case "task":
-                    printJson(runner.createFeedbackTask(required(runId, "run id"), required(feedbackId, "feedback id"), args.options));
-                    return;
-                case "resolve":
-                    printJson(runner.resolveFeedback(required(runId, "run id"), required(feedbackId, "feedback id"), args.options));
-                    return;
-                default:
-                    throw new Error("Usage: cw.js feedback list|show|summary|collect|task|resolve <run-id> [feedback-id]");
-            }
-        }
-        case "worker": {
-            const [subcommand, runId, workerId, resultPath] = args.positionals;
-            switch (subcommand) {
-                case "list":
-                    printJson(runner.listWorkers(required(runId, "run id"), args.options));
-                    return;
-                case "summary": {
-                    const summary = runner.summarizeWorkerRecords(required(runId, "run id"));
-                    if (wantsJson(args.options))
-                        printJson(summary);
-                    else
-                        process.stdout.write(`${(0, operator_ux_1.formatWorkerSummary)(summary)}\n`);
-                    return;
-                }
-                case "show":
-                    printJson(runner.showWorker(required(runId, "run id"), required(workerId, "worker id")));
-                    return;
-                case "manifest":
-                    printJson(runner.showWorkerManifest(required(runId, "run id"), required(workerId, "worker id")));
-                    return;
-                case "output":
-                    printJson(runner.recordWorkerOutput(required(runId, "run id"), required(workerId, "worker id"), required(resultPath, "result file"), args.options));
-                    return;
-                case "fail":
-                    printJson(runner.recordWorkerFailure(required(runId, "run id"), required(workerId, "worker id"), String(args.options.message || required(resultPath, "failure message")), args.options));
-                    return;
-                case "validate": {
-                    // Non-null = a boundary violation: a validate verb must report an invalid
-                    // verdict through its exit code, not just print it and exit 0.
-                    const violation = runner.validateWorker(required(runId, "run id"), required(workerId, "worker id"), resultPath);
-                    printJson(violation);
-                    if (violation)
-                        process.exitCode = 1;
-                    return;
-                }
-                default:
-                    throw new Error("Usage: cw.js worker list|summary|show|manifest|output|fail|validate <run-id> [worker-id] [result-file]");
-            }
-        }
-        case "audit": {
-            const [subcommand, runId, id] = args.positionals;
-            switch (subcommand) {
-                case "summary":
-                    printJson(runner.auditSummary(required(runId, "run id")));
-                    return;
-                case "verify": {
-                    const result = (0, capability_core_1.auditVerify)(runner, { ...args.options, runId: required(runId, "run id") });
-                    printJson(result);
-                    // Fail-closed: any unverified chain exits non-zero so `cw audit verify
-                    // <run> && deploy` stops — mirrors the telemetry-verify guard. verifyTrustAudit
-                    // returns verified:true for a truly absent/empty chain (nothing to prove),
-                    // so this stays exit 0 there; a FULLY-corrupt log reports present:false but
-                    // verified:false (corruptLines>0) and must NOT be conflated with absent — the
-                    // earlier `present && ...` guard let that severe tamper escape (exit 0).
-                    if (!result.verified)
-                        process.exitCode = 1;
-                    return;
-                }
-                case "worker":
-                    printJson(runner.workerAudit(required(runId, "run id"), required(id, "worker id")));
-                    return;
-                case "provenance":
-                    printJson(runner.evidenceProvenance(required(runId, "run id"), args.options));
-                    return;
-                case "multi-agent": {
-                    const view = runner.auditMultiAgent(required(runId, "run id"));
-                    if (wantsJson(args.options))
-                        printJson(view);
-                    else
-                        process.stdout.write(`${(0, operator_ux_1.formatMultiAgentTrustAudit)(view)}\n`);
-                    return;
-                }
-                case "policy": {
-                    const view = runner.auditPolicy(required(runId, "run id"));
-                    if (wantsJson(args.options))
-                        printJson(view);
-                    else
-                        process.stdout.write(`${(0, operator_ux_1.formatMultiAgentTrustAudit)(view)}\n`);
-                    return;
-                }
-                case "role": {
-                    const view = runner.auditRole(required(runId, "run id"), required(id, "role id"));
-                    if (wantsJson(args.options))
-                        printJson(view);
-                    else
-                        process.stdout.write(`${(0, operator_ux_1.formatMultiAgentTrustAudit)(view)}\n`);
-                    return;
-                }
-                case "blackboard": {
-                    const view = runner.auditBlackboard(required(runId, "run id"));
-                    if (wantsJson(args.options))
-                        printJson(view);
-                    else
-                        process.stdout.write(`${(0, operator_ux_1.formatMultiAgentTrustAudit)(view)}\n`);
-                    return;
-                }
-                case "judge": {
-                    const view = runner.auditJudge(required(runId, "run id"));
-                    if (wantsJson(args.options))
-                        printJson(view);
-                    else
-                        process.stdout.write(`${(0, operator_ux_1.formatMultiAgentTrustAudit)(view)}\n`);
-                    return;
-                }
-                case "attest":
-                    printJson(runner.recordAuditAttestation(required(runId, "run id"), args.options));
-                    return;
-                case "decision":
-                    printJson(runner.recordAuditDecision(required(runId, "run id"), required(id, "worker id"), args.options));
-                    return;
-                default:
-                    throw new Error("Usage: cw.js audit summary|worker|provenance|multi-agent|policy|role|blackboard|judge|attest|decision <run-id> [worker-id|role-id]");
-            }
-        }
-        case "candidate": {
-            const [subcommand, runId, candidateId, reason] = args.positionals;
-            switch (subcommand) {
-                case "list":
-                    printJson(runner.listCandidates(required(runId, "run id"), args.options));
-                    return;
-                case "show":
-                    printJson(runner.showCandidate(required(runId, "run id"), required(candidateId, "candidate id")));
-                    return;
-                case "register":
-                    printJson(runner.registerCandidate(required(runId, "run id"), args.options));
-                    return;
-                case "score":
-                    printJson(runner.scoreCandidate(required(runId, "run id"), required(candidateId, "candidate id"), args.options));
-                    return;
-                case "rank":
-                    printJson(runner.rankCandidates(required(runId, "run id"), args.options));
-                    return;
-                case "select":
-                    printJson(runner.selectCandidate(required(runId, "run id"), required(candidateId, "candidate id"), args.options));
-                    return;
-                case "reject":
-                    printJson(runner.rejectCandidate(required(runId, "run id"), required(candidateId, "candidate id"), String(args.options.reason || args.options.message || reason || "rejected")));
-                    return;
-                case "summary":
-                    if (wantsJson(args.options))
-                        printJson(runner.summarizeCandidateOperatorRecords(required(runId, "run id")));
-                    else
-                        process.stdout.write(`${(0, operator_ux_1.formatCandidateSummary)(runner.summarizeCandidateOperatorRecords(required(runId, "run id")))}\n`);
-                    return;
-                default:
-                    throw new Error("Usage: cw.js candidate list|show|register|score|rank|select|reject|summary <run-id> [candidate-id]");
-            }
-        }
         // ---- Team Collaboration (v0.1.32) ------------------------------------
-        case "approve": {
-            const [targetKind, runId, targetId] = args.positionals;
-            printJson(runner.collaborationApprove(required(runId, "run id"), required(targetKind, "target kind (candidate|commit|selection|run|task|node)"), required(targetId, "target id"), args.options));
+        case "approve":
+            (0, collaboration_1.handleApprove)(args, runner);
             return;
-        }
-        case "reject": {
-            const [targetKind, runId, targetId] = args.positionals;
-            printJson(runner.collaborationReject(required(runId, "run id"), required(targetKind, "target kind (candidate|commit|selection|run|task|node)"), required(targetId, "target id"), args.options));
+        case "reject":
+            (0, collaboration_1.handleReject)(args, runner);
             return;
-        }
-        case "comment": {
-            const [subcommand, ...rest] = args.positionals;
-            if (subcommand === "add") {
-                const [targetKind, runId, targetId] = rest;
-                printJson(runner.collaborationComment(required(runId, "run id"), required(targetKind, "target kind"), required(targetId, "target id"), args.options));
-                return;
-            }
-            if (subcommand === "list") {
-                const result = runner.collaborationCommentList(required(rest[0], "run id"), args.options);
-                if (wantsJson(args.options))
-                    printJson(result);
-                else
-                    process.stdout.write(`${runner.formatCommentList(result.comments)}\n`);
-                return;
-            }
-            throw new Error("Usage: cw.js comment add <kind> <run-id> <target-id> --body <text> | comment list <run-id> [--json]");
-        }
-        case "handoff": {
-            const [targetKind, runId, targetIdRaw] = args.positionals;
-            const kind = required(targetKind, "target kind (run|task|candidate|commit|node)");
-            const rid = required(runId, "run id");
-            const targetId = targetIdRaw || (kind === "run" ? rid : undefined);
-            printJson(runner.collaborationHandoff(rid, kind, required(targetId, "target id"), args.options));
+        case "comment":
+            (0, collaboration_1.handleComment)(args, runner);
+            return;
+        case "handoff":
+            (0, collaboration_1.handleHandoff)(args, runner);
+            return;
+        case "review":
+            (0, collaboration_1.handleReview)(args, runner);
             return;
-        }
-        case "review": {
-            const [subcommand, runId] = args.positionals;
-            if (subcommand === "status") {
-                const report = runner.reviewStatus(required(runId, "run id"), args.options);
-                if (wantsJson(args.options))
-                    printJson(report);
-                else
-                    process.stdout.write(`${runner.formatReviewStatus(report)}\n`);
-                return;
-            }
-            if (subcommand === "policy") {
-                printJson(runner.reviewPolicy(required(runId, "run id"), args.options));
-                return;
-            }
-            throw new Error("Usage: cw.js review status <run-id> [--json] | review policy <run-id> --required-approvals N --authorized-roles a,b --applies-to commit,selection");
-        }
         case "loop": {
-            printJson(scheduler.create({ ...args.options, kind: "loop" }));
+            (0, io_1.printJson)(scheduler.create({ ...args.options, kind: "loop" }));
             return;
         }
-        case "schedule": {
-            const [subcommand, id] = args.positionals;
-            switch (subcommand) {
-                case "create":
-                    printJson(scheduler.create(args.options));
-                    return;
-                case "list":
-                    printJson(scheduler.list(args.options.status ? String(args.options.status) : undefined));
-                    return;
-                case "delete":
-                    printJson(scheduler.delete(required(id, "schedule id")));
-                    return;
-                case "due":
-                    printJson(scheduler.due());
-                    return;
-                case "complete":
-                    printJson(scheduler.complete(required(id, "schedule id"), args.options));
-                    return;
-                case "pause":
-                    printJson(scheduler.pause(required(id, "schedule id")));
-                    return;
-                case "resume":
-                    printJson(scheduler.resume(required(id, "schedule id")));
-                    return;
-                case "run-now":
-                    printJson(scheduler.runNow(required(id, "schedule id")));
-                    return;
-                case "history":
-                    printJson(scheduler.history(id));
-                    return;
-                case "daemon": {
-                    const daemon = new daemon_1.DesktopSchedulerDaemon({
-                        cwd: String(args.options.cwd || process.cwd()),
-                        intervalSeconds: Number(args.options.intervalSeconds || args.options.interval || 60)
-                    });
-                    if (args.options.once) {
-                        printJson(daemon.tick());
-                        return;
-                    }
-                    await daemon.run();
-                    return;
-                }
-                default:
-                    throw new Error("Usage: cw.js schedule create|list|delete|due|complete|pause|resume|run-now|history|daemon");
-            }
-        }
-        case "routine": {
-            const [subcommand, idOrKind, payloadPath] = args.positionals;
-            switch (subcommand) {
-                case "create":
-                    printJson(triggers.create(args.options));
-                    return;
-                case "list":
-                    printJson(triggers.list(args.options.kind ? String(args.options.kind) : undefined));
-                    return;
-                case "delete":
-                    printJson(triggers.delete(required(idOrKind, "trigger id")));
-                    return;
-                case "fire": {
-                    const kind = required(idOrKind, "trigger kind");
-                    const payload = payloadPath ? JSON.parse(node_fs_1.default.readFileSync(payloadPath, "utf8")) : args.options;
-                    printJson(triggers.fire(kind, payload));
-                    return;
-                }
-                case "events":
-                    printJson(triggers.events(idOrKind));
-                    return;
-                default:
-                    throw new Error("Usage: cw.js routine create|list|delete|fire|events");
-            }
-        }
-        case "registry": {
-            const registry = (0, capability_core_1.runRegistryFor)(args.options, runner);
-            const [subcommand] = args.positionals;
-            switch (subcommand) {
-                case "refresh": {
-                    const report = (0, capability_core_1.runRegistryRefresh)(registry, args.options);
-                    if (wantsJson(args.options))
-                        printJson(report);
-                    else
-                        process.stdout.write(`${(0, run_registry_1.formatRegistryReport)(report)}\n`);
-                    return;
-                }
-                case "show": {
-                    const report = (0, capability_core_1.runRegistryShow)(registry, args.options);
-                    if (wantsJson(args.options))
-                        printJson(report);
-                    else
-                        process.stdout.write(`${(0, run_registry_1.formatRegistryReport)(report)}\n`);
-                    return;
-                }
-                default:
-                    throw new Error("Usage: cw.js registry refresh|show [--scope repo|home] [--json]");
-            }
-        }
-        case "metrics": {
-            const [subcommand, runId] = args.positionals;
-            switch (subcommand) {
-                case "show": {
-                    const report = runner.metricsShow(required(runId, "run id"), args.options);
-                    if (wantsJson(args.options))
-                        printJson(report);
-                    else
-                        process.stdout.write(`${(0, observability_1.formatMetricsReport)(report)}\n`);
-                    return;
-                }
-                case "summary": {
-                    const report = (0, capability_core_1.metricsSummary)((0, capability_core_1.runRegistryFor)(args.options, runner), runner, args.options);
-                    if (wantsJson(args.options))
-                        printJson(report);
-                    else
-                        process.stdout.write(`${(0, observability_1.formatMetricsSummary)(report)}\n`);
-                    return;
-                }
-                default:
-                    throw new Error("Usage: cw.js metrics show <run-id> | metrics summary [--scope repo|home] [--pricing <path>|default] [--json]");
-            }
-        }
-        case "run": {
-            // Agent Delegation Drive (v0.1.38): `cw run <app> --drive [--once]` drives a
-            // run end-to-end by delegating each worker to the agent backend. Distinct from
-            // the run-REGISTRY verbs below. `--preview` (or the `run drive <run-id>` form)
-            // is the read-only, deterministic next-step preview.
-            //
-            // A run-REGISTRY subcommand keyword (resume/show/...) must NOT be intercepted
-            // here just because it carries a --drive flag of its own — e.g.
-            // `run resume <id> --drive` is the resume verb's opt-in continuation, not
-            // `run <app=resume> --drive`. Fall through to the switch for those keywords.
-            const runRegistrySubcommand = new Set([
-                "drive", "search", "list", "show", "resume", "archive", "rerun", "export", "import", "verify-import", "inspect-archive"
-            ]);
-            if (args.options.drive && !runRegistrySubcommand.has(String(args.positionals[0] || ""))) {
-                const target = args.positionals[0];
-                const runId = optionalArg(args.options.run) || optionalArg(args.options.runId);
-                if (args.options.preview) {
-                    printJson((0, capability_core_1.runDrivePreview)(runner, { ...args.options, runId: runId || target }));
-                    return;
-                }
-                const driveArgs = { ...args.options };
-                if (runId)
-                    driveArgs.runId = runId;
-                else
-                    driveArgs.appId = target;
-                const dr = (0, capability_core_1.runDrive)(runner, driveArgs);
-                printJson(dr);
-                if (!wantsJson(args.options)) {
-                    emitRunSummary(runner, args.options, {
-                        runId: dr.runId,
-                        reportPath: dr.reportPath,
-                        status: dr.status,
-                        statePath: dr.statePath,
-                        completedWorkers: dr.completedWorkers,
-                        plannedWorkers: dr.plannedWorkers,
-                        agentConfigured: dr.agentConfigured
-                    });
-                }
-                return;
-            }
-            const registry = (0, capability_core_1.runRegistryFor)(args.options, runner);
-            const [subcommand, id] = args.positionals;
-            switch (subcommand) {
-                case "drive": {
-                    // `run drive <run-id>` = read-only preview; `--step [--once]` = mutating drive.
-                    if (args.options.step) {
-                        const driveArgs = { ...args.options };
-                        if (id)
-                            driveArgs.runId = id;
-                        const dr = (0, capability_core_1.runDrive)(runner, driveArgs);
-                        printJson(dr);
-                        if (!wantsJson(args.options)) {
-                            emitRunSummary(runner, args.options, {
-                                runId: dr.runId,
-                                reportPath: dr.reportPath,
-                                status: dr.status,
-                                statePath: dr.statePath,
-                                completedWorkers: dr.completedWorkers,
-                                plannedWorkers: dr.plannedWorkers,
-                                agentConfigured: dr.agentConfigured
-                            });
-                        }
-                        return;
-                    }
-                    printJson((0, capability_core_1.runDrivePreview)(runner, { ...args.options, runId: required(id, "run id") }));
-                    return;
-                }
-                case "search": {
-                    const result = (0, capability_core_1.runSearch)(registry, args.options);
-                    if (wantsJson(args.options))
-                        printJson(result);
-                    else
-                        process.stdout.write(`${(0, run_registry_1.formatRunSearch)(result)}\n`);
-                    return;
-                }
-                case "list": {
-                    const result = (0, capability_core_1.runList)(registry, args.options);
-                    if (wantsJson(args.options))
-                        printJson(result);
-                    else
-                        process.stdout.write(`${(0, run_registry_1.formatRunSearch)(result)}\n`);
-                    return;
-                }
-                case "show": {
-                    const result = (0, capability_core_1.runShow)(registry, required(id, "run id"), args.options);
-                    if (wantsJson(args.options))
-                        printJson(result);
-                    else
-                        process.stdout.write(`${(0, run_registry_1.formatRunShow)(result)}\n`);
-                    return;
-                }
-                case "resume": {
-                    const result = (0, capability_core_1.runResume)(registry, runner, required(id, "run id"), args.options);
-                    if (wantsJson(args.options))
-                        printJson(result);
-                    else
-                        process.stdout.write(`${(0, run_registry_1.formatResume)(result)}\n`);
-                    return;
-                }
-                case "archive":
-                    printJson((0, capability_core_1.runArchive)(registry, id, args.options));
-                    return;
-                case "rerun":
-                    printJson((0, capability_core_1.runRerun)(registry, required(id, "run id"), args.options));
-                    return;
-                case "export":
-                    printJson((0, capability_core_1.runExportArchive)(runner, required(id || optionalArg(args.options.runId || args.options.run), "run id"), args.options));
-                    return;
-                case "import":
-                    printJson((0, capability_core_1.runImportArchive)(runner, { ...args.options, archive: id || args.options.archive || args.options.path }));
-                    return;
-                case "verify-import": {
-                    const result = (0, capability_core_1.runVerifyImport)(runner, required(id || optionalArg(args.options.runId || args.options.run), "run id"), args.options);
-                    printJson(result);
-                    // Fail-closed ONLY behind --strict, so the default exit stays 0
-                    // (byte-identical). With --strict, any failed restore check — including
-                    // the new trust-audit row — exits 1 for `verify-import && restore`.
-                    if (Boolean(args.options.strict) && !result.ok)
-                        process.exitCode = 1;
-                    return;
-                }
-                case "inspect-archive": {
-                    const result = (0, capability_core_1.runInspectArchive)(runner, { ...args.options, archive: id || args.options.archive || args.options.path });
-                    printJson(result);
-                    // Read-only diagnostic: exit 1 when the archive fails any integrity check,
-                    // so `cw run inspect-archive <path> && restore` stops on a bad archive.
-                    if (!result.ok)
-                        process.exitCode = 1;
-                    return;
-                }
-                default:
-                    throw new Error("Usage: cw.js run search|list|show|resume|archive|rerun|drive|export|import|verify-import|inspect-archive [run-id|archive] [--scope repo|home] [--json]  |  cw.js run <app> --drive [--once] [--incremental] [--repo R --question Q]");
-            }
-        }
-        case "queue": {
-            const registry = (0, capability_core_1.runRegistryFor)(args.options, runner);
-            const [subcommand, id] = args.positionals;
-            switch (subcommand) {
-                case "add":
-                    printJson((0, capability_core_1.queueAdd)(registry, args.options));
-                    return;
-                case "list": {
-                    const result = (0, capability_core_1.queueList)(registry, args.options);
-                    if (wantsJson(args.options))
-                        printJson(result);
-                    else
-                        process.stdout.write(`${(0, run_registry_1.formatQueueList)(result)}\n`);
-                    return;
-                }
-                case "drain":
-                    printJson((0, capability_core_1.queueDrain)(registry, args.options));
-                    return;
-                case "show":
-                    printJson((0, capability_core_1.queueShow)(registry, required(id, "queue id")));
-                    return;
-                default:
-                    throw new Error("Usage: cw.js queue add|list|drain|show [queue-id] [--repo PATH] [--priority N]");
-            }
-        }
-        case "sched": {
-            const registry = (0, capability_core_1.runRegistryFor)(args.options, runner);
-            const [subcommand, idArg] = args.positionals;
-            switch (subcommand) {
-                case "plan":
-                    printJson((0, capability_core_1.schedPlan)(registry, args.options));
-                    return;
-                case "lease":
-                    printJson((0, capability_core_1.schedLease)(registry, args.options));
-                    return;
-                case "release":
-                    printJson((0, capability_core_1.schedRelease)(registry, { ...args.options, leaseId: args.options.leaseId || idArg }));
-                    return;
-                case "complete":
-                    printJson((0, capability_core_1.schedComplete)(registry, { ...args.options, leaseId: args.options.leaseId || idArg }));
-                    return;
-                case "reclaim":
-                    printJson((0, capability_core_1.schedReclaim)(registry, args.options));
-                    return;
-                case "reset":
-                    printJson((0, capability_core_1.schedReset)(registry, { ...args.options, id: args.options.id || idArg }));
-                    return;
-                case "policy": {
-                    const [, action] = args.positionals;
-                    if (action === "set") {
-                        printJson((0, capability_core_1.schedPolicySet)(registry, args.options));
-                        return;
-                    }
-                    printJson((0, capability_core_1.schedPolicyShow)(registry));
-                    return;
-                }
-                default:
-                    throw new Error("Usage: cw.js sched plan|lease|release|complete|reclaim|reset|policy [show|set] [id] [--maxConcurrent N --maxAttempts N ...]");
-            }
-        }
-        case "clones": {
-            // Remote-source clone cache (v0.1.91): `list` inspects the ~/.local/state/cool-workflow
-            // /clones checkouts that `--link`/URL reviews populate; `gc` reclaims them (a TTL sweep,
-            // or --all). Pure filesystem work — no network, no run registry.
-            const [subcommand] = args.positionals;
-            switch (subcommand) {
-                case "list": {
-                    const result = (0, capability_core_1.listClones)(args.options);
-                    if (wantsJson(args.options))
-                        printJson(result);
-                    else
-                        process.stdout.write(`${formatClonesList(result)}\n`);
-                    return;
-                }
-                case "gc": {
-                    const result = (0, capability_core_1.gcClones)(args.options);
-                    if (wantsJson(args.options))
-                        printJson(result);
-                    else
-                        process.stdout.write(`${formatClonesGc(result)}\n`);
-                    return;
-                }
-                default:
-                    throw new Error("Usage: cw.js clones list [--json] | clones gc [--older-than-days N] [--all] [--json]");
-            }
-        }
-        case "gc": {
-            // Run Retention & Provable Reclamation (v0.1.39). `plan` is a pure dry-run
-            // (frees nothing); `run` executes the write-ahead reclamation transaction;
-            // `verify` re-proves a reclaimed run. CW never reclaims by default.
-            const registry = (0, capability_core_1.runRegistryFor)(args.options, runner);
-            const [subcommand, id] = args.positionals;
-            switch (subcommand) {
-                case "plan": {
-                    const result = (0, capability_core_1.gcPlan)(registry, id, args.options);
-                    if (wantsJson(args.options))
-                        printJson(result);
-                    else
-                        process.stdout.write(`${(0, run_registry_1.formatGcPlan)(result)}\n`);
-                    return;
-                }
-                case "run": {
-                    const result = (0, capability_core_1.gcRun)(registry, id, args.options);
-                    if (wantsJson(args.options))
-                        printJson(result);
-                    else
-                        process.stdout.write(`${(0, run_registry_1.formatGcRun)(result)}\n`);
-                    return;
-                }
-                case "verify": {
-                    const result = (0, capability_core_1.gcVerify)(registry, required(id, "run id"), args.options);
-                    if (wantsJson(args.options))
-                        printJson(result);
-                    else
-                        process.stdout.write(`${(0, run_registry_1.formatGcVerify)(result)}\n`);
-                    // Fail closed ONLY on a real integrity failure: a run that WAS reclaimed
-                    // but no longer re-proves. A not-reclaimed run has nothing to verify
-                    // (reclaimed:false/verified:false) and must not be treated as a failure.
-                    // LIMIT (honest): a DELETED reclaimed.json reads as reclaimed:false, so
-                    // proof-deletion is indistinguishable from never-reclaimed here without
-                    // an independent witness (e.g. a trust-audit reclamation event) — a
-                    // follow-up. This guard is still strictly better than the prior exit-0.
-                    if (result.reclaimed && !result.verified)
-                        process.exitCode = 1;
-                    return;
-                }
-                default:
-                    throw new Error("Usage: cw.js gc plan|run|verify [run-id] [--reclaimAfterArchiveDays N] [--keep-scratch] [--keep-snapshots] [--limit N] [--json]");
-            }
-        }
-        case "history": {
-            const registry = (0, capability_core_1.runRegistryFor)(args.options, runner);
-            const result = (0, capability_core_1.runHistory)(registry, args.options);
-            if (wantsJson(args.options))
-                printJson(result);
-            else
-                process.stdout.write(`${(0, run_registry_1.formatHistory)(result)}\n`);
+        case "schedule":
+            await (0, scheduling_1.handleSchedule)(args, scheduler);
+            return;
+        case "routine":
+            (0, scheduling_1.handleRoutine)(args, triggers);
+            return;
+        case "registry":
+            (0, registry_1.handleRegistry)(args, runner);
+            return;
+        case "metrics":
+            (0, operational_1.handleMetrics)(args, runner);
+            return;
+        case "run":
+            (0, run_1.handleRun)(args, runner);
+            return;
+        case "queue":
+            (0, registry_1.handleQueue)(args, runner);
+            return;
+        case "sched":
+            (0, scheduling_1.handleSched)(args, runner);
+            return;
+        case "clones":
+            (0, clones_1.handleClones)(args);
+            return;
+        case "gc":
+            (0, maintenance_1.handleGc)(args, runner);
+            return;
+        case "history":
+            (0, registry_1.handleHistory)(args, runner);
+            return;
+        case "telemetry":
+            (0, maintenance_1.handleTelemetry)(args, runner);
+            return;
+        case "demo":
+            (0, maintenance_1.handleDemo)(args, runner);
+            return;
+        case "workbench":
+            await (0, workbench_1.handleWorkbench)(args, runner);
             return;
-        }
-        case "telemetry": {
-            const [subcommand, id] = args.positionals;
-            switch (subcommand) {
-                case "verify": {
-                    const result = (0, capability_core_1.telemetryVerify)(runner, { ...args.options, runId: id || args.options.runId || args.options.run });
-                    if (wantsJson(args.options))
-                        printJson(result);
-                    else
-                        process.stdout.write(`${(0, telemetry_demo_1.formatTelemetryVerify)(result)}\n`);
-                    // Fail closed: a forged/edited/corrupt ledger verifies false — report it
-                    // through the exit code so `cw telemetry verify <run> && deploy` cannot
-                    // pass on a lie. (Absent ledger = present:false/verified:true -> exit 0.)
-                    if (!result.verified)
-                        process.exitCode = 1;
-                    return;
-                }
-                default:
-                    throw new Error("Usage: cw.js telemetry verify <run-id> [--pubkey <pem-or-path>] [--json]");
-            }
-        }
-        case "demo": {
-            const [subcommand] = args.positionals;
-            switch (subcommand) {
-                case "tamper": {
-                    const result = (0, capability_core_1.demoTamper)(runner, args.options);
-                    if (wantsJson(args.options))
-                        printJson(result);
-                    else
-                        process.stdout.write(`${(0, telemetry_demo_1.formatTamperDemo)(result)}\n`);
-                    // Fail closed: if the proof did not hold (a tamper went undetected),
-                    // exit nonzero so the demo can never green a broken guarantee.
-                    if (!result.proven)
-                        process.exitCode = 1;
-                    return;
-                }
-                case "bundle": {
-                    const result = (0, capability_core_1.demoBundle)(runner, args.options);
-                    if (wantsJson(args.options))
-                        printJson(result);
-                    else
-                        process.stdout.write(`${(0, telemetry_demo_1.formatBundleDemo)(result)}\n`);
-                    // Fail closed: a forged bundle that verified would be a regression in the
-                    // bundle guarantee — exit nonzero so the demo can never green it.
-                    if (!result.proven)
-                        process.exitCode = 1;
-                    return;
-                }
-                default:
-                    throw new Error("Usage: cw.js demo tamper|bundle [--json]");
-            }
-        }
-        case "workbench": {
-            const [subcommand, runId] = args.positionals;
-            switch (subcommand) {
-                case "view": {
-                    // Read-only five-panel view of one run. Same core entry as cw_workbench_view.
-                    const view = (0, workbench_1.buildWorkbenchRunView)(runner, required(runId, "run id"));
-                    if (wantsJson(args.options))
-                        printJson(view);
-                    else
-                        process.stdout.write(`${formatWorkbenchView(view)}\n`);
-                    return;
-                }
-                case "serve": {
-                    // The OPTIONAL localhost host. `--once`/`--json` emit the descriptor only
-                    // (no server); the default starts the read-only, localhost-only host.
-                    if (args.options.once || wantsJson(args.options)) {
-                        printJson((0, workbench_1.buildWorkbenchServeDescriptor)(runner, { ...args.options, once: true }));
-                        return;
-                    }
-                    const host = new workbench_host_1.WorkbenchHost({
-                        runner,
-                        cwd: String(args.options.cwd || process.cwd()),
-                        port: Number(args.options.port) || undefined,
-                        scope: args.options.scope === "repo" ? "repo" : "home"
-                    });
-                    await host.run();
-                    return;
-                }
-                default:
-                    throw new Error("Usage: cw.js workbench serve [--port N] [--once] | view <run-id> [--json]");
-            }
-        }
         default:
             throw new Error(`Unknown command: ${args.command}${((0, orchestrator_1.suggestCommand)(String(args.command || "")) ? `. Did you mean: ${(0, orchestrator_1.suggestCommand)(String(args.command))}?` : "")}`);
     }
 }
-function required(value, label) {
-    if (!value)
-        throw new Error(`Missing ${label}.\n  Tip: find run ids with "cw run list" or create one with "cw quickstart"`);
-    return value;
-}
-function optionalArg(value) {
-    return typeof value === "string" && value.trim() ? value.trim() : undefined;
-}
-/** Emit the calm end-of-run summary (stderr, TTY-gated inside the reporter): the COMPACT findings
- *  table re-parsed from each completed worker's `cw:result`, the report path, where the per-worker
- *  transcripts live, and — under `--full` — the report inline. Stderr/human-side ONLY: stdout (the
- *  `--json` payload printed just before this) stays byte-exact. Shared by the quickstart and the
- *  two `run --drive` paths so all three render an identical summary. */
-function emitRunSummary(runner, options, fields) {
-    // Anchor run reads to the run's OWN repo (a drive/quickstart may run cross-directory): the run
-    // dir is <repo>/.cw/runs/<id>/, holding each worker's transcript.md next to its result.md.
-    const runDir = typeof fields.statePath === "string" ? node_path_1.default.dirname(fields.statePath) : undefined;
-    const baseDir = runDir ? node_path_1.default.resolve(runDir, "..", "..", "..") : undefined;
-    const findings = (0, capability_core_1.collectRunFindings)(runner, fields.runId, baseDir);
-    // --full ALSO prints the report inline at run end (the compact table stays the default summary).
-    let fullReport;
-    if (options.full && fields.reportPath && node_fs_1.default.existsSync(fields.reportPath)) {
-        try {
-            fullReport = node_fs_1.default.readFileSync(fields.reportPath, "utf8");
-        }
-        catch { /* best-effort inline */ }
-    }
-    reporter_1.reporter.runSummary({
-        runId: fields.runId,
-        reportPath: fields.reportPath,
-        status: fields.status,
-        completedWorkers: fields.completedWorkers,
-        plannedWorkers: fields.plannedWorkers,
-        agentConfigured: fields.agentConfigured,
-        findings,
-        runDir,
-        fullReport
-    });
-}
-function printJson(value) {
-    process.stdout.write(`${JSON.stringify(value, null, 2)}\n`);
-}
-function humanBytes(n) {
-    if (n < 1024)
-        return `${n}B`;
-    const units = ["KiB", "MiB", "GiB"];
-    let v = n / 1024;
-    let i = 0;
-    while (v >= 1024 && i < units.length - 1) {
-        v /= 1024;
-        i += 1;
-    }
-    return `${v.toFixed(1)}${units[i]}`;
-}
-function formatClonesList(result) {
-    if (result.count === 0)
-        return `No cached remote checkouts in ${result.clonesDir}.`;
-    const rows = result.entries.map((e) => {
-        const when = e.fetchedAt ? e.fetchedAt.replace("T", " ").replace(/\..*$/, "Z") : "unknown";
-        return `  ${e.kind.padEnd(7)} ${humanBytes(e.bytes).padStart(8)}  ${when}  ${e.url}${e.ref ? `@${e.ref}` : ""}`;
-    });
-    return [
-        `${result.count} cached checkout${result.count === 1 ? "" : "s"} — ${humanBytes(result.totalBytes)} in ${result.clonesDir}`,
-        "  KIND       SIZE  FETCHED               SOURCE",
-        ...rows,
-        `\nReclaim with: cw clones gc --older-than-days 30   (or --all)`
-    ].join("\n");
-}
-function formatClonesGc(result) {
-    const scope = result.all ? "all entries" : `entries older than ${result.olderThanDays} day(s)`;
-    if (result.removed.length === 0)
-        return `Nothing to reclaim (${scope}); ${result.keptCount} kept in ${result.clonesDir}.`;
-    const rows = result.removed.map((r) => `  ${humanBytes(r.bytes).padStart(8)}  ${r.url}`);
-    return [
-        `Reclaimed ${result.removed.length} checkout${result.removed.length === 1 ? "" : "s"} (${scope}) — freed ${humanBytes(result.freedBytes)}; ${result.keptCount} kept`,
-        ...rows
-    ].join("\n");
-}
-function wantsJson(options) {
-    return Boolean(options.json || options.format === "json");
-}
 /** Prompt the user for a question interactively when --question is missing on a TTY. */
 async function promptQuestion(options) {
     if (options.question || !process.stdin.isTTY)
@@ -1659,17 +486,3 @@ async function promptQuestion(options) {
         });
     });
 }
-function formatWorkbenchView(view) {
-    const lines = [
-        `Workbench view ${view.runId} (${view.resolved ? "resolved" : "UNRESOLVED"})`,
-        view.error ? `  error: ${view.error}` : ""
-    ].filter(Boolean);
-    for (const [group, panels] of Object.entries(view.panels)) {
-        lines.push(`  ${group}:`);
-        for (const [name, panel] of Object.entries(panels)) {
-            const note = panel.status === "present" ? panel.capability : `absent (${panel.error || "unreadable"})`;
-            lines.push(`    ${name}: ${panel.status} — ${note}`);
-        }
-    }
-    return lines.join("\n");
-}