cool-workflow 0.1.98 → 0.2.1

package/dist/shell/pipeline-cli.js

@@ -0,0 +1,667 @@
+"use strict";
+// shell/pipeline-cli.ts — CLI-facing entry points for the pipeline/drive
+// spine: planRun, runDrivePreview, runDriveStep, quickstartRun,
+// dispatchRun, recordResultRun, commitRun.
+//
+// MILESTONE 6+7 (combined). Wires the pieces built in shell/pipeline.ts,
+// shell/drive.ts, shell/dispatch.ts, shell/commit.ts, shell/worker-
+// isolation.ts, shell/workflow-app-loader.ts into the shapes
+// core/capability-table.ts's CLI bindings call.
+//
+// Evidence: SPEC/pipeline-run.md "CLI / MCP surface (capability layer)".
+var __createBinding = (this && this.__createBinding) || (Object.create ? (function(o, m, k, k2) {
+    if (k2 === undefined) k2 = k;
+    var desc = Object.getOwnPropertyDescriptor(m, k);
+    if (!desc || ("get" in desc ? !m.__esModule : desc.writable || desc.configurable)) {
+      desc = { enumerable: true, get: function() { return m[k]; } };
+    }
+    Object.defineProperty(o, k2, desc);
+}) : (function(o, m, k, k2) {
+    if (k2 === undefined) k2 = k;
+    o[k2] = m[k];
+}));
+var __setModuleDefault = (this && this.__setModuleDefault) || (Object.create ? (function(o, v) {
+    Object.defineProperty(o, "default", { enumerable: true, value: v });
+}) : function(o, v) {
+    o["default"] = v;
+});
+var __importStar = (this && this.__importStar) || (function () {
+    var ownKeys = function(o) {
+        ownKeys = Object.getOwnPropertyNames || function (o) {
+            var ar = [];
+            for (var k in o) if (Object.prototype.hasOwnProperty.call(o, k)) ar[ar.length] = k;
+            return ar;
+        };
+        return ownKeys(o);
+    };
+    return function (mod) {
+        if (mod && mod.__esModule) return mod;
+        var result = {};
+        if (mod != null) for (var k = ownKeys(mod), i = 0; i < k.length; i++) if (k[i] !== "default") __createBinding(result, mod, k[i]);
+        __setModuleDefault(result, mod);
+        return result;
+    };
+})();
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.planRun = planRun;
+exports.runDrivePreview = runDrivePreview;
+exports.runDriveStep = runDriveStep;
+exports.quickstartRun = quickstartRun;
+exports.dispatchRun = dispatchRun;
+exports.recordResultRun = recordResultRun;
+exports.commitRun = commitRun;
+const fs = __importStar(require("node:fs"));
+const path = __importStar(require("node:path"));
+const pipeline_1 = require("./pipeline");
+const workflow_app_loader_1 = require("./workflow-app-loader");
+const drive_1 = require("./drive");
+const dispatch_1 = require("./dispatch");
+const commit_1 = require("./commit");
+const worker_isolation_1 = require("./worker-isolation");
+const observability_1 = require("./observability");
+const run_store_1 = require("./run-store");
+const report_1 = require("./report");
+const agent_config_1 = require("./agent-config");
+const remote_source_1 = require("./remote-source");
+const trust_audit_1 = require("./trust-audit");
+const report_cli_1 = require("./report-cli");
+const QUICKSTART_DEFAULT_APP = "architecture-review";
+/** True when `value` is a truthy CLI flag (present, or "true"/"1"/"yes"/"on"). */
+function truthyFlag(value) {
+    if (value === true)
+        return true;
+    if (typeof value === "string")
+        return /^(1|true|yes|on)$/i.test(value.trim());
+    return false;
+}
+/** First non-empty string among the given arg values (mirrors the old
+ *  build's optionalString read over several alias keys). */
+function firstString(...values) {
+    for (const value of values) {
+        if (typeof value === "string" && value.trim())
+            return value.trim();
+    }
+    return undefined;
+}
+/** Shell-quote a token for a copy-pasteable next command (byte-exact to the
+ *  old build's shellWord). */
+function shellWord(value) {
+    if (/^[A-Za-z0-9_./:@%+=,-]+$/.test(value))
+        return value;
+    return `'${value.replace(/'/g, "'\\''")}'`;
+}
+/** The copy-pasteable `cw quickstart …` line the `--check` payload echoes,
+ *  weaving in --bundle / --with-trust-key / --strict-signatures so the
+ *  suggested resume preserves bundle intent. Byte-exact port of the old
+ *  build's quickstartNextCommand. */
+function quickstartNextCommand(appId, repo, args) {
+    const parts = ["cw", "quickstart", shellWord(appId), "--repo", shellWord(repo)];
+    const question = firstString(args.question);
+    if (question)
+        parts.push("--question", shellWord(question));
+    const command = firstString(args.agentCommand, args["agent-command"]);
+    if (command)
+        parts.push("--agent-command", shellWord(command));
+    if (truthyFlag(args.bundle))
+        parts.push("--bundle");
+    const trustKey = firstString(args["with-trust-key"], args.withTrustKey, args.trustKey);
+    if (trustKey)
+        parts.push("--with-trust-key", shellWord(trustKey));
+    if (truthyFlag(args["strict-signatures"]) || truthyFlag(args.strictSignatures) || truthyFlag(args.strictSigs))
+        parts.push("--strict-signatures");
+    return parts.join(" ");
+}
+/** Runtime keys that must NEVER leak into run.inputs (they are drive/CLI
+ *  plumbing, not workflow-declared inputs). Byte-exact to the old
+ *  build's DRIVE_RUNTIME_KEYS list. */
+const RUNTIME_KEYS = new Set([
+    "once", "now", "preview", "step", "drive", "json", "format", "run", "runId", "cwd",
+    "agentCommand", "agent-command", "agentArgs", "agent-args", "agentEndpoint", "agent-endpoint",
+    "agentModel", "agent-model", "agentTimeoutMs", "agent-timeout-ms", "resume", "incremental",
+    "concurrency", "link", "ref", "branch", "refresh", "check", "app", "appId", "workflowId", "question", "repo",
+]);
+/** Byte-exact port of the old build's `normalizeInputs`
+ *  (src/orchestrator/lifecycle-operations.ts:465-480): repeated `--arg
+ *  key=value` pairs unpack into inputs (key = text before the first "=",
+ *  value = the rest re-joined with "="); `repo` copies to `cwd` when `cwd`
+ *  is not already set. Per SPEC/orchestrator.md's "Plan input rules" and
+ *  SPEC/workflow-apps.md. */
+function planInputsFor(args) {
+    const out = {};
+    for (const [key, value] of Object.entries(args)) {
+        if (key === "arg") {
+            const pairs = Array.isArray(value) ? value : [value];
+            for (const pair of pairs) {
+                const [argKey, ...rest] = String(pair).split("=");
+                out[argKey] = rest.join("=");
+            }
+            continue;
+        }
+        if (RUNTIME_KEYS.has(key))
+            continue;
+        out[key] = value;
+    }
+    if (typeof args.repo === "string")
+        out.repo = args.repo;
+    if (typeof args.question === "string")
+        out.question = args.question;
+    // An explicit --cwd is stripped by RUNTIME_KEYS above, but the old build
+    // honored it for the run anchor. Re-add it (like repo) so a caller-supplied
+    // cwd is not silently dropped to process.cwd() — a cross-request bleed.
+    if (typeof args.cwd === "string" && args.cwd.trim())
+        out.cwd = args.cwd;
+    if (out.repo && !out.cwd)
+        out.cwd = out.repo;
+    return out;
+}
+function invocationCwd(args) {
+    return typeof args.cwd === "string" && args.cwd.trim() ? path.resolve(args.cwd) : process.cwd();
+}
+/** True for a value that counts as "not supplied" — byte-exact to the old
+ *  build's cli-options.isMissing (undefined / null / empty string). */
+function isMissingInput(value) {
+    return value === undefined || value === null || value === "";
+}
+/** Resolve a workflow app for `plan`/`run --drive` over the SAME surface `cw
+ *  list` shows — bundled apps AND legacy `<name>.workflow.js` files. The old
+ *  build's plan() resolved via loadWorkflowAppById (full discovery over both the
+ *  workflows/ and apps/ roots — app-operations.ts:64), so any id `cw list`
+ *  surfaces is plannable. v2's fast-path loadWorkflowApp only reads
+ *  apps/<id>/app.json, so a legacy workflow-file wrapper (e.g.
+ *  legacy-research-synthesis) that `list` CAN resolve died with "Workflow app
+ *  not found". Fall back to the full-discovery record (loadWorkflowAppRecordById)
+ *  and adapt it to the minimal LoadedWorkflowApp plan() consumes. */
+function resolveWorkflowAppForPlan(appId) {
+    try {
+        return (0, workflow_app_loader_1.loadWorkflowApp)(appId);
+    }
+    catch (error) {
+        if (!(error instanceof workflow_app_loader_1.WorkflowAppNotFoundError))
+            throw error;
+        const record = (0, workflow_app_loader_1.loadWorkflowAppRecordById)(appId);
+        const author = typeof record.app.author === "string" ? record.app.author : record.app.author?.name;
+        return {
+            id: record.app.id,
+            title: record.app.title,
+            summary: record.app.summary || record.app.workflow.summary || "",
+            version: record.app.version,
+            ...(author !== undefined ? { author } : {}),
+            workflow: record.app.workflow,
+            sandboxProfiles: record.app.sandboxProfiles || record.app.workflow.sandboxProfiles || [],
+            sourcePath: record.source.manifestPath || record.source.path,
+        };
+    }
+}
+/** `cw plan <workflowId>` — real: loads the app, plans a fresh run,
+ *  returns the canonical plan summary. */
+function planRun(args) {
+    const appId = String(args.workflowId || args.app || QUICKSTART_DEFAULT_APP);
+    // POLA: `cw plan <app>` does NOT default repo to the caller's cwd (unlike the
+    // one-command quickstart). The old build validated required inputs FIRST
+    // (lifecycle-operations.ts validateInputs, message `Missing required input
+    // --<name>`), so a missing --repo surfaces the copy-pasteable `-dir` recovery
+    // line (cli/entry.ts recoveryHint's "missing"+"repo" branch) instead of a
+    // silent cwd-anchored run. Auto-filling repo here would hide that recovery.
+    const app = resolveWorkflowAppForPlan(appId);
+    const planInputs = planInputsFor(args);
+    for (const declared of app.workflow.inputs || []) {
+        if (declared.required && isMissingInput(planInputs[declared.name])) {
+            throw new Error(`Missing required input --${declared.name}`);
+        }
+    }
+    const run = (0, pipeline_1.plan)(app, planInputs);
+    // `pendingTasks` is the canonical plan-payload key both `cw plan` and
+    // `cw_plan` carry (old build src/capability-core.ts:79 planSummary:
+    // `pendingTasks: run.tasks.filter(status === "pending").length`, and
+    // SPEC/workflow-apps.md). `taskCount` stays as a harmless extra.
+    const pendingTasks = run.tasks.filter((task) => task.status === "pending").length;
+    return { schemaVersion: 1, runId: run.id, workflowId: run.workflow.id, statePath: run.paths.state, reportPath: run.paths.report, pendingTasks, taskCount: run.tasks.length };
+}
+function runDrivePreview(args) {
+    const runId = String(args.runId || args.run || "");
+    const cwd = invocationCwd(args);
+    return (0, drive_1.drivePreview)(runId, cwd, args);
+}
+/** `cw run <app|--run id> --drive [--once]` — plans a fresh run (unless
+ *  `--run` continues an existing one) and drives it. */
+function runDriveStep(args) {
+    const existingRunId = String(args.runId || args.run || "");
+    const options = {
+        once: Boolean(args.once),
+        now: typeof args.now === "string" ? args.now : undefined,
+        args,
+        concurrency: args.concurrency !== undefined ? Number(args.concurrency) : undefined,
+        incremental: Boolean(args.incremental),
+    };
+    if (existingRunId) {
+        const cwd = invocationCwd(args);
+        const run = (0, run_store_1.loadRunFromCwd)(existingRunId, cwd);
+        return (0, drive_1.drive)(existingRunId, run.cwd, options);
+    }
+    const appId = String(args.appId || args.app || args.positionalApp || "");
+    if (!appId)
+        throw new Error("run --drive requires an app id (or --run <run-id> to continue)");
+    if (!args.repo && !args.cwd)
+        args.repo = invocationCwd(args);
+    const app = (0, workflow_app_loader_1.loadWorkflowApp)(appId);
+    const run = (0, pipeline_1.plan)(app, planInputsFor(args));
+    return (0, drive_1.drive)(run.id, run.cwd, options);
+}
+/** `cw quickstart [app] --check` — read-only preflight: does the app
+ *  resolve, is the repo readable/writable, is a question set, is an
+ *  agent backend configured. Never plans or writes a run. Byte-exact
+ *  port of the old build's `quickstartCheck` (src/capability-core.ts),
+ *  local-repo path only (the --link/remote preflight variant is not
+ *  ported — no conformance case exercises it). */
+function quickstartCheck(appId, args, remoteCandidate) {
+    // `--link`/URL preflight: validate the URL SHAPE + tooling WITHOUT fetching
+    // (a clone is heavy + side-effecting; --check stays read-only). Swaps the
+    // local-repo readability checks for link + tooling. `repo` carries the
+    // sanitized URL so the result reports what would be reviewed.
+    if (remoteCandidate)
+        return remoteQuickstartCheck(appId, args, remoteCandidate);
+    const base = invocationCwd(args);
+    const repoArg = typeof args.repo === "string" && args.repo.trim() ? args.repo : base;
+    const repo = path.resolve(base, repoArg);
+    const checks = [];
+    try {
+        (0, workflow_app_loader_1.showWorkflowApp)(appId);
+        checks.push({ name: "app", status: "ok", detail: `Workflow app ${appId} is available.` });
+    }
+    catch {
+        checks.push({
+            name: "app",
+            status: "blocked",
+            detail: `Workflow app ${appId} is not available.`,
+            fix: "Run `cw app list` and choose one of the listed app ids.",
+        });
+    }
+    let repoReadable = false;
+    let repoStateWritable = false;
+    try {
+        const stat = fs.statSync(repo);
+        repoReadable = stat.isDirectory();
+        if (!repoReadable)
+            throw new Error("not a directory");
+        fs.accessSync(repo, fs.constants.R_OK);
+        checks.push({ name: "repo", status: "ok", detail: `Repository path is readable (${repo}).` });
+    }
+    catch {
+        checks.push({
+            name: "repo",
+            status: "blocked",
+            detail: `Repository path is not readable (${repo}).`,
+            fix: "Pass --repo PATH for a readable repository directory.",
+        });
+    }
+    try {
+        const cwDir = path.join(repo, ".cw");
+        fs.accessSync(fs.existsSync(cwDir) ? cwDir : repo, fs.constants.W_OK);
+        repoStateWritable = repoReadable;
+        checks.push({ name: "repo-state", status: "ok", detail: "Run state location is writable." });
+    }
+    catch {
+        checks.push({
+            name: "repo-state",
+            status: "blocked",
+            detail: "Run state location is not writable.",
+            fix: "Use a writable repo, fix directory permissions, or pass --repo to a writable checkout.",
+        });
+    }
+    if (typeof args.question === "string" && args.question.trim()) {
+        checks.push({ name: "question", status: "ok", detail: "Question is set." });
+    }
+    else {
+        checks.push({ name: "question", status: "blocked", detail: "Question is missing.", fix: "Pass --question TEXT." });
+    }
+    if ((0, agent_config_1.agentConfigured)(args)) {
+        checks.push({ name: "agent", status: "ok", detail: "Agent backend is configured." });
+    }
+    else {
+        checks.push({
+            name: "agent",
+            status: "blocked",
+            detail: "No agent backend is configured.",
+            fix: 'Pass --agent-command "claude -p", set $CW_AGENT_COMMAND, or use --agent-command builtin:claude.',
+        });
+    }
+    // --bundle preflight: a completed run sealed into a bundle re-verifies offline
+    // only with a public trust key. Warn (not block) by default; block only under
+    // --strict-signatures where an unkeyed bundle would fail verification.
+    if (truthyFlag(args.bundle)) {
+        const trustKey = firstString(args["with-trust-key"], args.withTrustKey, args.trustKey, args.pubkey) || process.env.CW_AGENT_ATTEST_PUBKEY;
+        if (trustKey) {
+            checks.push({ name: "bundle-trust-key", status: "ok", detail: "Bundle trust public key is configured." });
+        }
+        else if (truthyFlag(args["strict-signatures"]) || truthyFlag(args.strictSignatures) || truthyFlag(args.strictSigs)) {
+            checks.push({ name: "bundle-trust-key", status: "blocked", detail: "Strict signature verification needs a public trust key.", fix: "Pass --with-trust-key PATH or set $CW_AGENT_ATTEST_PUBKEY." });
+        }
+        else {
+            checks.push({ name: "bundle-trust-key", status: "warn", detail: "No public trust key is configured; unsigned or unkeyed bundles may verify with reduced signature proof.", fix: "Pass --with-trust-key PATH to embed the public key." });
+        }
+    }
+    const ok = checks.every((check) => check.status !== "blocked") && repoStateWritable;
+    return { schemaVersion: 1, mode: "check", ok, appId, repo, checks, nextCommand: quickstartNextCommand(appId, repo, args) };
+}
+/** `--check` for a `--link`/URL review: validates the URL shape + git tooling
+ *  WITHOUT fetching (byte-behavior port of the old build's remoteQuickstartCheck).
+ *  `repo` carries the sanitized URL. */
+function remoteQuickstartCheck(appId, args, candidate) {
+    const validation = (0, remote_source_1.validateRemoteUrl)(candidate);
+    const checks = [];
+    try {
+        (0, workflow_app_loader_1.showWorkflowApp)(appId);
+        checks.push({ name: "app", status: "ok", detail: `Workflow app ${appId} is available.` });
+    }
+    catch {
+        checks.push({ name: "app", status: "blocked", detail: `Workflow app ${appId} is not available.`, fix: "Run `cw app list` and choose one of the listed app ids." });
+    }
+    if (validation.ok) {
+        checks.push({ name: "link", status: "ok", detail: `Remote source is a valid ${validation.kind} URL (${validation.url}).` });
+    }
+    else {
+        checks.push({ name: "link", status: "blocked", detail: `Remote source is not usable: ${validation.reason}.`, fix: "Pass a git URL (https/ssh/git/file or git@host:repo)." });
+    }
+    if ((0, remote_source_1.gitAvailable)()) {
+        checks.push({ name: "tooling", status: "ok", detail: "git is available to clone the remote." });
+    }
+    else {
+        checks.push({ name: "tooling", status: "blocked", detail: "git was not found on PATH.", fix: "Install git, then re-run." });
+    }
+    if (typeof args.question === "string" && args.question.trim()) {
+        checks.push({ name: "question", status: "ok", detail: "Question is set." });
+    }
+    else {
+        checks.push({ name: "question", status: "blocked", detail: "Question is missing.", fix: "Pass --question TEXT." });
+    }
+    if ((0, agent_config_1.agentConfigured)(args)) {
+        checks.push({ name: "agent", status: "ok", detail: "Agent backend is configured." });
+    }
+    else {
+        checks.push({ name: "agent", status: "blocked", detail: "No agent backend is configured.", fix: 'Pass --agent-command "claude -p", set $CW_AGENT_COMMAND, or use --agent-command builtin:claude.' });
+    }
+    const ok = checks.every((check) => check.status !== "blocked");
+    const question = firstString(args.question);
+    const nextCommand = `cw quickstart ${shellWord(appId)} --link ${shellWord(validation.url)}${question ? ` --question ${shellWord(question)}` : ""}`;
+    return { schemaVersion: 1, mode: "check", ok, appId, repo: validation.url, checks, nextCommand };
+}
+/** `cw quickstart [app] --question ...` — composes plan -> runDrive ->
+ *  report in one call. Default app is architecture-review. `--check` is a
+ *  read-only preflight that never plans/drives/writes (see
+ *  `quickstartCheck` above). `--preview` is a read-only next-step
+ *  projection (never drives), `--resume` advances one step (no --run) or
+ *  continues a named run to completion (--run <id>) — both ported byte-for-
+ *  byte from the old build's src/capability-core.ts quickstart(). */
+function quickstartRun(args) {
+    const appId = String(args.appId || args.app || args.workflowId || QUICKSTART_DEFAULT_APP);
+    // Remote source: a `--link <url>` — or a URL passed to `--repo`/`-dir` — is
+    // materialized to a LOCAL checkout HERE (capability/shell layer). Cloning is
+    // non-deterministic network I/O and must never enter the replay-deterministic
+    // core, so we rewrite `args.repo`/`args.cwd` to the local path; everything
+    // downstream is a normal local run.
+    const linkArg = typeof args.link === "string" && args.link.trim() ? args.link.trim() : undefined;
+    const repoArgRaw = typeof args.repo === "string" && args.repo.trim() ? args.repo.trim() : undefined;
+    const remoteCandidate = linkArg || (repoArgRaw && (0, remote_source_1.isRemoteUrl)(repoArgRaw) ? repoArgRaw : undefined);
+    if (!remoteCandidate && !args.repo && !args.cwd)
+        args.repo = invocationCwd(args);
+    if (Boolean(args.check))
+        return quickstartCheck(appId, args, remoteCandidate);
+    // Materialize the remote NOW — after `--check` (never fetches) and before any
+    // plan/drive — so the core only ever sees the local checkout. Fails closed: a
+    // bad URL / blocked scheme / missing git / fetch failure throws here.
+    let remoteSource;
+    if (remoteCandidate) {
+        remoteSource = (0, remote_source_1.materializeRemote)(remoteCandidate, {
+            ref: typeof args.ref === "string" ? args.ref : typeof args.branch === "string" ? args.branch : undefined,
+            refresh: Boolean(args.refresh),
+        });
+        args.repo = remoteSource.localPath;
+        args.cwd = remoteSource.localPath;
+        // Record the origin as plan INPUTS so it rides into run.inputs → the report
+        // header (report.ts renders `- Source: url@sha` from run.inputs.sourceUrl).
+        args.sourceUrl = remoteSource.url;
+        args.sourceCommit = remoteSource.commit;
+        if (remoteSource.ref)
+            args.sourceRef = remoteSource.ref;
+    }
+    // `--resume`: a discoverability flag over the existing continuation. With no
+    // `--run`, advance exactly ONE step (reuse the `--once` path) and print a
+    // copy-pasteable continue line; with `--run <id>`, continue that run to
+    // completion (the default drive). It adds no new execution path. Byte-exact to
+    // the old build's src/capability-core.ts quickstart().
+    const resume = Boolean(args.resume);
+    const existingRunId = String(args.runId || args.run || "");
+    const resumeRunId = resume && existingRunId ? existingRunId : undefined;
+    // `--preview`: read-only, deterministic next-step projection (no spawn, no
+    // commit). Plan a fresh run (the read-only first verb) then project the next
+    // drive step. Never drives.
+    if (Boolean(args.preview)) {
+        let previewRunId = existingRunId;
+        let repoCwd = typeof args.cwd === "string" && args.cwd.trim() ? args.cwd : typeof args.repo === "string" ? args.repo : undefined;
+        if (!previewRunId) {
+            const run = resolveWorkflowAppForPlan(appId);
+            const planned = (0, pipeline_1.plan)(run, planInputsFor(args));
+            previewRunId = planned.id;
+            repoCwd = planned.cwd;
+        }
+        const target = repoCwd && fs.existsSync(repoCwd) ? repoCwd : invocationCwd(args);
+        return (0, drive_1.drivePreview)(previewRunId, target, args);
+    }
+    const options = {
+        // `--resume` with no run id advances a SINGLE step (reuse `--once`), so a
+        // newcomer WITNESSES the stop-then-resume; `--resume --run <id>` continues to
+        // completion (the default drive). Non-resume paths keep the caller's --once.
+        once: Boolean(args.once) || (resume && !resumeRunId),
+        now: typeof args.now === "string" ? args.now : undefined,
+        args,
+        concurrency: args.concurrency !== undefined ? Number(args.concurrency) : undefined,
+        incremental: Boolean(args.incremental),
+    };
+    let run;
+    if (existingRunId) {
+        run = (0, run_store_1.loadRunFromCwd)(existingRunId, invocationCwd(args));
+    }
+    else {
+        run = (0, pipeline_1.plan)(resolveWorkflowAppForPlan(appId), planInputsFor(args));
+    }
+    const result = (0, drive_1.drive)(run.id, run.cwd, options);
+    const finalRun = (0, run_store_1.loadRunFromCwd)(run.id, run.cwd);
+    (0, report_1.writeReport)(finalRun);
+    // Tamper-evident provenance: bind the remote origin (url@sha) into the run's
+    // hash-chained trust-audit log so `cw audit verify` re-proves where the code
+    // came from. Best-effort — the origin is already in run.inputs/report/result.
+    if (remoteSource) {
+        try {
+            (0, trust_audit_1.recordTrustAuditEvent)(finalRun, {
+                kind: remoteSource.kind === "archive" ? "source.download" : "source.clone",
+                decision: "recorded",
+                source: "operator-recorded",
+                metadata: { url: remoteSource.url, commit: remoteSource.commit, ref: remoteSource.ref || null, kind: remoteSource.kind, depth: 1 },
+            });
+        }
+        catch {
+            /* provenance is additive; never fail a completed review over an audit hiccup */
+        }
+    }
+    // --bundle: after a COMPLETE drive, seal the run into a portable, self-verified
+    // bundle so the one command yields a client-verifiable artifact. Pure composition
+    // of reportBundleCli (export sealed + offline self-verify); spawns nothing. Gated
+    // on completion: a partial/blocked run is NEVER sealed. Run-state resolution
+    // anchors to the run's OWN repo (run.cwd) — quickstart runs cross-directory — but
+    // OUTPUT paths resolve against the CALLER's cwd so artifacts land where the
+    // operator ran the command. Byte-behavior port of the old build's quickstart().
+    const wantsBundle = truthyFlag(args.bundle);
+    let bundle;
+    if (wantsBundle && result.status === "complete") {
+        const callerBase = invocationCwd(args);
+        const outArg = firstString(args.output, args.path, args.archive);
+        const extractArg = firstString(args["extract-report"], args.extractReport, args.extractReportTo);
+        bundle = (0, report_cli_1.reportBundleCli)(result.runId, {
+            ...args,
+            cwd: run.cwd,
+            output: path.resolve(callerBase, outArg || `${result.runId}.cwrun.json`),
+            ...(extractArg ? { "extract-report": path.resolve(callerBase, extractArg) } : {}),
+        });
+    }
+    // Human-facing triage `hint` (stderr-side; absent on a clean completion so the
+    // default payload is byte-identical). Byte-exact wording to the old build's
+    // src/capability-core.ts quickstart(): the fail-closed "not configured …
+    // DELEGATES" line reaffirms the red line; the resume/once lines are copy-paste
+    // continue commands.
+    let hint;
+    if (!result.agentConfigured) {
+        hint =
+            "agent backend not configured — set CW_AGENT_COMMAND (e.g. \"claude -p\") or pass --agent-command, then re-run. The one command DELEGATES worker execution to YOUR agent; it never executes a model itself.";
+    }
+    else if (result.status === "parked") {
+        hint = `a worker parked past its retry budget — inspect: cw run show ${result.runId}`;
+    }
+    else if (result.status === "blocked") {
+        hint = `the drive is blocked — inspect: cw run drive ${result.runId}`;
+    }
+    else if (result.status === "in-progress") {
+        hint = resume
+            ? `one step advanced — continue: cw quickstart ${appId} --run ${result.runId} --resume${wantsBundle ? " --bundle" : ""}`
+            : `one step advanced (--once) — continue: cw quickstart ${appId} --run ${result.runId} --once`;
+    }
+    // --bundle on a run that did not complete is a NO-OP, not silence: tell the
+    // operator why nothing was sealed (the Rule of Silence permits a human hint).
+    if (wantsBundle && result.status !== "complete") {
+        hint = `${hint ? `${hint} ` : ""}--bundle skipped: the run did not complete (status=${result.status}); no bundle was sealed.`;
+    }
+    // Byte-exact to the old build's quickstart() return shape
+    // (src/capability-core.ts): `appId` is the resolved app id (the
+    // argument, or its architecture-review default), distinct from
+    // `workflowId` which is the driven run's own workflow id (equal for a
+    // top-level run, different for a sub-workflow hop). `remote` is present only
+    // for a --link/URL source, so a local-repo run stays byte-identical.
+    // `resumedFrom` is stamped ONLY when an explicit --run was continued
+    // (conditional spread keeps the key absent on the fresh/default path).
+    return {
+        appId,
+        ...result,
+        hint,
+        ...(resumeRunId ? { resumedFrom: resumeRunId } : {}),
+        // `bundle` is present only when --bundle sealed a completed run (conditional
+        // spread keeps the key absent on the default path → byte-identical output).
+        ...(bundle ? { bundle } : {}),
+        ...(remoteSource
+            ? { remote: { url: remoteSource.url, commit: remoteSource.commit, kind: remoteSource.kind, cached: remoteSource.cached, ...(remoteSource.ref ? { ref: remoteSource.ref } : {}) } }
+            : {}),
+    };
+}
+function dispatchRun(args) {
+    const runId = String(args.runId);
+    const run = (0, run_store_1.loadRunFromCwd)(runId, invocationCwd(args));
+    // parseArgv keys long flags in kebab-case; accept camelCase as a fallback.
+    const flag = (kebab, camel) => {
+        const v = args[kebab] ?? args[camel];
+        return typeof v === "string" && v.trim() ? v : undefined;
+    };
+    const manifest = (0, dispatch_1.createDispatchManifest)(run, args.limit !== undefined ? Number(args.limit) : undefined, {
+        sandboxProfileId: typeof args.sandbox === "string" ? args.sandbox : undefined,
+        sandbox: typeof args.sandbox === "string" ? args.sandbox : undefined,
+        backendId: typeof args.backend === "string" ? args.backend : undefined,
+        multiAgentRunId: flag("multi-agent-run", "multiAgentRun"),
+        multiAgentGroupId: flag("multi-agent-group", "multiAgentGroup"),
+        multiAgentRoleId: flag("multi-agent-role", "multiAgentRole"),
+        multiAgentFanoutId: flag("multi-agent-fanout", "multiAgentFanout"),
+    });
+    if (manifest.dispatchId) {
+        (0, commit_1.commitState)(run, `dispatch:${manifest.dispatchId}`);
+        (0, run_store_1.saveCheckpoint)(run);
+        (0, report_1.writeReport)(run);
+    }
+    return manifest;
+}
+function recordResultRun(args) {
+    const runId = String(args.runId);
+    const taskId = String(args.taskId);
+    const resultPath = String(args.resultPath);
+    const run = (0, run_store_1.loadRunFromCwd)(runId, invocationCwd(args));
+    const task = run.tasks.find((t) => t.id === taskId);
+    if (!task || !task.workerId)
+        throw new Error(`Unknown task id for run ${runId}: ${taskId}`);
+    const absolute = path.resolve(resultPath);
+    // A result path inside a system directory is never accepted (POLA): the
+    // operator file gets copied into the worker's result.md, so a /etc/passwd
+    // source would smuggle system content into a run. Byte-behavior port of the
+    // old build's recordResult system-directory blacklist.
+    if (/^\/(etc|bin|sbin|usr|Library|System|Applications|boot|dev|proc|sys|root|var\/log|var\/run)\//.test(absolute)) {
+        throw new Error(`Result path must not be a system directory: ${resultPath}`);
+    }
+    if (!fs.existsSync(absolute))
+        throw new Error(`Result file does not exist: ${resultPath}`);
+    const workerId = String(task.workerId);
+    // Host-attested `cw result <run> <task> <file>` intake: the operator hands CW
+    // an EXTERNAL result file that lives OUTSIDE the worker's read-only write
+    // boundary. The old task-level recordResult (lifecycle-operations.ts:279-280)
+    // COPIED that external file into the run's results area and recorded the
+    // internal path — it never ran the external path through validateSandboxWrite.
+    // v2 collapsed the two intakes into recordWorkerOutput, which sandbox-validates
+    // its input against the worker boundary, so a bare external path is rejected
+    // ("write path is outside sandbox profile <id>"). Restore the copy-in: stage
+    // the operator file at the worker's OWN result.md (which IS inside the write
+    // boundary), then record that internal path exactly like a driven worker.
+    const manifest = (0, worker_isolation_1.showWorkerManifest)(run, workerId);
+    fs.mkdirSync(path.dirname(manifest.resultPath), { recursive: true });
+    fs.copyFileSync(absolute, manifest.resultPath);
+    const output = (0, worker_isolation_1.recordWorkerOutput)(run, workerId, manifest.resultPath);
+    // Host-attested token usage (v0.1.31): record it verbatim as provenance when
+    // the operator supplied `--usage-*` flags; CW never synthesizes usage. The old
+    // task-level recordResult set `task.usage = usage` (lifecycle-operations.ts:286)
+    // and its unit was the TASK. v2 records through recordWorkerOutput, which gives
+    // the worker an `output` record — so the observability usage UNIT becomes the
+    // WORKER (deriveUsageTotals reads worker.usage for workers with output, and
+    // EXCLUDES that task). Attach the usage to the worker scope so the report counts
+    // it as an attested unit; also stamp task.usage for byte-parity with the old
+    // task-level record.
+    const usage = (0, observability_1.parseUsageFromArgs)(args, new Date().toISOString());
+    if (usage) {
+        task.usage = usage;
+        const scope = (0, worker_isolation_1.getWorkerScope)(run, workerId);
+        if (scope)
+            scope.usage = usage;
+    }
+    // Byte-exact to the old build's orchestrator recordWorkerOutput()
+    // wrapper: an accepted result is its own checkpoint commit, not just a
+    // bare saveCheckpoint (SPEC/pipeline-run.md's persist-ordering rule).
+    (0, commit_1.commitState)(run, `worker:${workerId}:result`);
+    (0, run_store_1.saveCheckpoint)(run);
+    (0, report_1.writeReport)(run);
+    return output;
+}
+/** `cw commit <run-id>` — byte-exact port of the old build's
+ *  `orchestrator/lifecycle-operations.ts`'s `commit()`: the CLI/MCP
+ *  payload wraps the commit record as `{runId, commit}` (NOT the commit
+ *  record at top level). Both the success AND the throw path write the
+ *  report + checkpoint before returning/re-throwing — a gate failure
+ *  still leaves the run's report/state current on disk. */
+function commitRun(args) {
+    const runId = String(args.runId);
+    const run = (0, run_store_1.loadRunFromCwd)(runId, invocationCwd(args));
+    const allowCheckpoint = Boolean(args.allowUnverifiedCheckpoint || args["allow-unverified-checkpoint"]);
+    const hasGateOption = Boolean(args.verifier || args.verifierNode || args["verifier-node"] || args.candidate || args.selection);
+    try {
+        const commit = (0, commit_1.commitState)(run, {
+            reason: typeof args.reason === "string" && args.reason ? args.reason : "manual",
+            verifierNodeId: (typeof args.verifier === "string" && args.verifier) ||
+                (typeof args.verifierNode === "string" && args.verifierNode) ||
+                (typeof args["verifier-node"] === "string" && args["verifier-node"]) ||
+                undefined,
+            candidateId: typeof args.candidate === "string" ? args.candidate : undefined,
+            selectionId: typeof args.selection === "string" ? args.selection : undefined,
+            verifierGated: hasGateOption || !allowCheckpoint,
+            allowUnverifiedCheckpoint: allowCheckpoint,
+            source: "cli",
+        });
+        (0, report_1.writeReport)(run);
+        (0, run_store_1.saveCheckpoint)(run);
+        return { runId: run.id, commit };
+    }
+    catch (error) {
+        (0, report_1.writeReport)(run);
+        (0, run_store_1.saveCheckpoint)(run);
+        throw error;
+    }
+}