feat: make chat-style tasks the default experience

packages/adapter-utils/src/server-utils.test.ts

@@ -887,6 +887,9 @@ describe("renderPaperclipWakePrompt", () => {
     expect(DEFAULT_PAPERCLIP_AGENT_PROMPT_TEMPLATE).not.toContain(
       "for request_confirmation this resumes only after acceptance",
     );
+    expect(DEFAULT_PAPERCLIP_AGENT_PROMPT_TEMPLATE).toContain(
+      "Never create probe or throwaway issue-thread interactions to discover the interactions API shape or your permissions",
+    );
     expect(DEFAULT_PAPERCLIP_AGENT_PROMPT_TEMPLATE).toContain("confirmation:{issueId}:plan:{revisionId}");
     expect(DEFAULT_PAPERCLIP_AGENT_PROMPT_TEMPLATE).toContain("Wait for acceptance before creating implementation subtasks");
     expect(DEFAULT_PAPERCLIP_AGENT_PROMPT_TEMPLATE).toContain(

packages/adapter-utils/src/server-utils.ts

@@ -170,6 +170,7 @@ export const DEFAULT_PAPERCLIP_AGENT_PROMPT_TEMPLATE = [
   "- Create child issues directly when you know what needs to be done; use issue-thread interactions when the board/user must choose suggested tasks, answer structured questions, or confirm a proposal.",
   "- Use `PAPERCLIP_SCRATCH_DIR` / `PAPERCLIP_RUN_SCRATCH_DIR` for temporary scratch files instead of ad hoc `/tmp` paths; Paperclip removes that run-owned directory after the run ends.",
   "- To ask for that input, create an interaction on the current issue with POST /api/issues/{issueId}/interactions using kind suggest_tasks, ask_user_questions, or request_confirmation. Use continuationPolicy wake_assignee when you need to resume after a response (it wakes on acceptance and rejection alike; only expiry does not wake); use wake_assignee_on_accept when you want to resume only after acceptance.",
+  "- Never create probe or throwaway issue-thread interactions to discover the interactions API shape or your permissions; schema discovery goes through the OpenAPI spec and explicit validation errors, not placeholder cards. Every ask_user_questions, suggest_tasks, or request_confirmation you post must carry a real, answerable prompt; withdraw one you no longer need instead of leaving it pending.",
   "- When you intentionally restart follow-up work on a completed assigned issue, include structured `resume: true` with the POST /api/issues/{issueId}/comments or PATCH /api/issues/{issueId} comment payload. Generic agent comments on closed issues are inert by default.",
   "- For plan approval, update the plan document first, then create request_confirmation targeting the latest plan revision with idempotencyKey confirmation:{issueId}:plan:{revisionId}. Wait for acceptance before creating implementation subtasks, and create a fresh confirmation after superseding board/user comments if approval is still needed.",
   "- If blocked, mark the issue blocked and name the unblock owner and action.",

packages/adapters/claude-local/src/server/acp.test.ts

@@ -85,7 +85,14 @@ afterEach(async () => {
     if (value === undefined) delete process.env[key];
     else process.env[key] = value;
   }
-  await Promise.all(tempRoots.splice(0).map((root) => fs.rm(root, { recursive: true, force: true })));
+  // The sandbox process-session bridge writes event files asynchronously; on slow
+  // CI shards a final write can race the recursive rm (ENOTEMPTY on the events
+  // dir), so let fs.rm retry until the writer has quiesced.
+  await Promise.all(
+    tempRoots
+      .splice(0)
+      .map((root) => fs.rm(root, { recursive: true, force: true, maxRetries: 10, retryDelay: 200 })),
+  );
 });
 
 class FakeRuntime {

packages/shared/src/constants.ts

@@ -295,6 +295,10 @@ export type IssueThreadInteractionContinuationPolicy =
 
 export const TASK_WATCHDOG_PRODUCT_BUG_ORIGIN_KIND = "task_watchdog_product_bug";
 
+// Marks the single onboarding "first task" so surfaces can special-case it
+// (e.g. suppress the seeded-description bubble and rely on a seeded greeting).
+export const ONBOARDING_FIRST_TASK_ORIGIN_KIND = "onboarding_first_task";
+
 export const ISSUE_ORIGIN_KINDS = [
   "manual",
   "routine_execution",
@@ -304,6 +308,7 @@ export const ISSUE_ORIGIN_KINDS = [
   "stranded_issue_recovery",
   "task_watchdog",
   TASK_WATCHDOG_PRODUCT_BUG_ORIGIN_KIND,
+  ONBOARDING_FIRST_TASK_ORIGIN_KIND,
 ] as const;
 export type BuiltInIssueOriginKind = (typeof ISSUE_ORIGIN_KINDS)[number];
 export type PluginIssueOriginKind = `plugin:${string}`;

packages/shared/src/feature-catalog.ts

@@ -96,14 +96,6 @@ export const INSTANCE_FEATURE_CATALOG: Record<InstanceFeatureKey, FeatureCatalog
     cloudDefault: false,
     selfHostedDefault: false,
   },
-  enableTaskChatRedesign: {
-    title: "Chat-Style Tasks",
-    description:
-      "Reimagines the task detail page as a live conversation with your agents: chat bubbles for people and agents, streaming activity — thinking, tool calls, diffs — that folds into a one-line summary when a turn finishes, inline plan/question/permission cards, a three-mode composer (Agent · Plan · Ask), and a resizable Properties · Plan · Artifacts pane.",
-    tier: "preference",
-    cloudDefault: false,
-    selfHostedDefault: false,
-  },
   enableTaskWatchdogs: {
     title: "Task Watchdogs",
     description:

packages/shared/src/index.ts

@@ -289,6 +289,7 @@ export {
   ISSUE_THREAD_INTERACTION_CONTINUATION_POLICIES,
   ISSUE_ORIGIN_KINDS,
   TASK_WATCHDOG_PRODUCT_BUG_ORIGIN_KIND,
+  ONBOARDING_FIRST_TASK_ORIGIN_KIND,
   ISSUE_WATCHDOG_DISCOVERY_KINDS,
   ISSUE_SURFACE_VISIBILITIES,
   ISSUE_RECOVERY_ACTION_KINDS,

packages/shared/src/types/instance.ts

@@ -52,7 +52,6 @@ export interface InstanceExperimentalSettings {
   enablePipelines: boolean;
   enableCases: boolean;
   enableConferenceRoomChat: boolean;
-  enableTaskChatRedesign: boolean;
   enableTaskWatchdogs: boolean;
   enableIssuePlanDecompositions: boolean;
   enableExperimentalFileViewer: boolean;

packages/shared/src/types/issue.ts

@@ -1077,6 +1077,14 @@ export interface AskUserQuestionsQuestionOption {
   id: string;
   label: string;
   description?: string | null;
+  /**
+   * When true, selecting this option reveals an inline free-text field instead
+   * of acting as an inert choice. The typed answer is submitted as the
+   * question's `otherText`. Author at most one free-text option per question and
+   * do not add dead "I'll describe it" options that only duplicate the built-in
+   * free-text affordance.
+   */
+  freeText?: boolean;
 }
 
 export interface AskUserQuestionsQuestion {
@@ -1109,8 +1117,11 @@ export interface AskUserQuestionsResult {
   answers: AskUserQuestionsAnswer[];
   cancelled?: true;
   cancellationReason?: string | null;
-  expirationReason?: "superseded_by_comment";
+  expirationReason?: "superseded_by_comment" | "superseded_by_newer_interaction";
   commentId?: string | null;
+  // Set with expirationReason "superseded_by_newer_interaction": the newer
+  // sibling ask_user_questions that replaced this one (PAP-437).
+  supersededByInteractionId?: string | null;
   summaryMarkdown?: string | null;
 }
 

packages/shared/src/validators/instance.ts

@@ -46,7 +46,6 @@ export const instanceExperimentalSettingsSchema = z.object({
   enablePipelines: z.boolean().default(false),
   enableCases: z.boolean().default(false),
   enableConferenceRoomChat: z.boolean().default(false),
-  enableTaskChatRedesign: z.boolean().default(false),
   enableTaskWatchdogs: z.boolean().default(false),
   enableIssuePlanDecompositions: z.boolean().default(false),
   enableExperimentalFileViewer: z.boolean().default(false),

packages/shared/src/validators/issue.ts

@@ -491,13 +491,24 @@ const createIssueDuplicateGuardSchema = {
     .default(false),
 };
 
+// Narrow intent flag set by the onboarding wizard on the single first task. The
+// server owns the resulting origin kind (clients cannot set arbitrary origin
+// kinds) and uses it to seed the agent greeting instead of an LLM welcome step.
+const onboardingFirstTaskMarkerSchema = {
+  onboardingFirstTask: z.boolean().optional(),
+};
+
 export const createIssueInputSchema = createIssueBaseSchema.extend({
   status: createIssueBaseSchema.shape.status.optional(),
   ...createIssueDuplicateGuardSchema,
+  ...onboardingFirstTaskMarkerSchema,
 });
 
 export const createIssueSchema = withCreateIssueStatusDefault(
-  createIssueBaseSchema.extend(createIssueDuplicateGuardSchema),
+  createIssueBaseSchema.extend({
+    ...createIssueDuplicateGuardSchema,
+    ...onboardingFirstTaskMarkerSchema,
+  }),
 ).superRefine(requireBlockedStatusForUnblockDescriptor);
 
 export type CreateIssue = z.infer<typeof createIssueSchema>;
@@ -759,6 +770,12 @@ export const askUserQuestionsQuestionOptionSchema = z.object({
   id: z.string().trim().min(1).max(120),
   label: z.string().trim().min(1).max(120),
   description: z.string().trim().max(500).nullable().optional(),
+  freeText: z
+    .boolean()
+    .optional()
+    .describe(
+      "When true, selecting this option reveals an inline text field; the typed value is returned as the question's otherText. Use this for a real \"I'll describe it\" choice instead of authoring a dead option that does nothing. At most one free-text option per question.",
+    ),
 });
 
 export const askUserQuestionsQuestionSchema = z.object({
@@ -789,6 +806,7 @@ export const askUserQuestionsPayloadSchema = z.object({
     seenQuestionIds.add(question.id);
 
     const seenOptionIds = new Set<string>();
+    let freeTextOptionCount = 0;
     for (const [optionIndex, option] of question.options.entries()) {
       if (seenOptionIds.has(option.id)) {
         ctx.addIssue({
@@ -798,6 +816,16 @@ export const askUserQuestionsPayloadSchema = z.object({
         });
       }
       seenOptionIds.add(option.id);
+      if (option.freeText) {
+        freeTextOptionCount += 1;
+        if (freeTextOptionCount > 1) {
+          ctx.addIssue({
+            code: z.ZodIssueCode.custom,
+            message: "A question may declare at most one free-text option",
+            path: ["questions", questionIndex, "options", optionIndex, "freeText"],
+          });
+        }
+      }
     }
   }
 });
@@ -815,8 +843,11 @@ export const askUserQuestionsResultSchema = z.object({
   answers: z.array(askUserQuestionsAnswerSchema).max(20),
   cancelled: z.literal(true).optional(),
   cancellationReason: z.string().trim().max(4000).nullable().optional(),
-  expirationReason: z.literal("superseded_by_comment").optional(),
+  expirationReason: z.enum(["superseded_by_comment", "superseded_by_newer_interaction"]).optional(),
   commentId: z.string().uuid().nullable().optional(),
+  // Set alongside expirationReason "superseded_by_newer_interaction": the id of
+  // the newer sibling ask_user_questions that replaced this one (PAP-437).
+  supersededByInteractionId: z.string().uuid().nullable().optional(),
   summaryMarkdown: z.string().max(20000).nullable().optional(),
 });
 

server/src/__tests__/built-in-agents.test.ts

@@ -569,7 +569,7 @@ describeEmbeddedPostgres("built-in agents", () => {
     });
   });
 
-  it("auto-provisions a paused Reflection Coach bundle with skill sync and a disabled routine", async () => {
+  it("reconciles an enabled Reflection Coach bundle with skill sync and a disabled routine", async () => {
     const companyId = await seedCompany({ requireApproval: false });
     const root = await agentService(db).create(companyId, {
       name: "CEO",
@@ -581,6 +581,17 @@ describeEmbeddedPostgres("built-in agents", () => {
       permissions: {},
     });
 
+    // The Reflection Coach is opt-in (not auto-created). Enabling it on demand
+    // materializes its managed bundle in a single pass.
+    const enabled = await builtInAgentService(db).ensure(companyId, "reflection-coach");
+    expect(enabled.agent?.adapterConfig).toMatchObject({
+      instructionsBundleMode: "managed",
+      instructionsEntryFile: "AGENTS.md",
+    });
+    expect(enabled.agent?.adapterConfig).not.toMatchObject({ model: "gpt-5.4", apiKey: "do-not-copy" });
+
+    // Startup reconcile keeps the enabled bundle tracking stock and re-grants
+    // the root/company default permissions.
     const result = await reconcileBuiltInAgentsOnStartup(db);
     expect(result.autoEnsured).toBeGreaterThanOrEqual(1);
     expect(result.defaultGrantsEnsured).toBeGreaterThanOrEqual(4);
@@ -606,11 +617,6 @@ describeEmbeddedPostgres("built-in agents", () => {
         },
       },
     });
-    expect(state.agent?.adapterConfig).toMatchObject({
-      instructionsBundleMode: "managed",
-      instructionsEntryFile: "AGENTS.md",
-    });
-    expect(state.agent?.adapterConfig).not.toMatchObject({ model: "gpt-5.4", apiKey: "do-not-copy" });
     expect(state.resources.map((resource) => [resource.resourceKind, resource.stockStatus])).toEqual([
       ["instructions", "stock_current"],
       ["skill", "stock_current"],
@@ -693,7 +699,7 @@ describeEmbeddedPostgres("built-in agents", () => {
     )).size).toBe(3);
   });
 
-  it("preserves new-agent approval gates during automatic Reflection Coach provisioning", async () => {
+  it("preserves new-agent approval gates during on-demand Reflection Coach provisioning", async () => {
     const companyId = await seedCompany({ requireApproval: true });
     const root = await agentService(db).create(companyId, {
       name: "CEO",
@@ -710,12 +716,11 @@ describeEmbeddedPostgres("built-in agents", () => {
       applyInSeparateFollowUpRun: true,
     };
 
-    const result = await reconcileBuiltInAgentsOnStartup(db);
-
-    expect(result).toMatchObject({
-      autoEnsured: 2,
-      pendingApprovals: 2,
-    });
+    // The Reflection Coach is opt-in, so it's enabled on demand. With board
+    // approval required, provisioning it must leave a pending agent + a
+    // hire_agent approval rather than an active agent.
+    const provisioned = await builtInAgentService(db).provision(companyId, "reflection-coach");
+    expect(provisioned.approval).not.toBeNull();
     const state = await builtInAgentService(db).get(companyId, "reflection-coach");
     expect(state).toMatchObject({
       status: "pending_approval",
@@ -751,7 +756,7 @@ describeEmbeddedPostgres("built-in agents", () => {
     });
 
     const pendingReconcile = await reconcileBuiltInAgentsOnStartup(db);
-    expect(pendingReconcile.pendingApprovals).toBe(2);
+    expect(pendingReconcile.pendingApprovals).toBe(1);
     const stillPending = await builtInAgentService(db).get(companyId, "reflection-coach");
     expect(stillPending).toMatchObject({
       status: "pending_approval",
@@ -787,7 +792,7 @@ describeEmbeddedPostgres("built-in agents", () => {
     const agentRows = await db.select().from(agents).where(eq(agents.companyId, companyId));
     expect(agentRows.filter((row) => readBuiltInAgentMarker(row.metadata)?.key === "reflection-coach")).toHaveLength(1);
     const approvalRows = await db.select().from(approvals).where(eq(approvals.companyId, companyId));
-    expect(approvalRows).toHaveLength(2);
+    expect(approvalRows).toHaveLength(1);
   });
 
   it("preserves Reflection Coach instruction drift on reconcile and restores it on reset", async () => {
@@ -1062,7 +1067,21 @@ describeEmbeddedPostgres("built-in agents", () => {
     const { olderId, newerId } = await seedLegacyDuplicateBriefs(affectedCompanyId);
     // A second company created after the affected one — previously skipped
     // entirely because the duplicate error escaped the reconciliation loop.
+    // Give it a drifted built-in row so we can prove reconcile still reached it.
     const healthyCompanyId = await seedCompany({ requireApproval: false });
+    const healthyBriefsId = randomUUID();
+    await db.insert(agents).values({
+      id: healthyBriefsId,
+      companyId: healthyCompanyId,
+      name: "Stale Briefs",
+      role: "engineer",
+      status: "idle",
+      adapterType: "codex_local",
+      adapterConfig: { model: "gpt-5.4" },
+      runtimeConfig: {},
+      permissions: {},
+      metadata: withBuiltInAgentMarker({}, { key: "briefs", featureKeys: ["briefs"] }),
+    });
 
     const result = await reconcileBuiltInAgentsOnStartup(db);
     expect(result.companyFailures).toBe(0);
@@ -1077,9 +1096,10 @@ describeEmbeddedPostgres("built-in agents", () => {
       ),
     ).toHaveLength(1);
 
-    // The company after the affected one still had its bundled agents provisioned.
-    const healthyCoach = await builtInAgentService(db).get(healthyCompanyId, "reflection-coach");
-    expect(healthyCoach.agentId).toBeTruthy();
+    // The company after the affected one was still reconciled (its drifted
+    // built-in row was repaired to stock) rather than skipped.
+    const [healthyBriefs] = await db.select().from(agents).where(eq(agents.id, healthyBriefsId));
+    expect(healthyBriefs?.name).toBe("Briefs Agent");
   });
 
   it("automatically materializes the Reflection Coach bundle without enabling background work", async () => {

server/src/__tests__/companies-service.test.ts

@@ -24,7 +24,7 @@ import {
 } from "./helpers/embedded-postgres.js";
 import { companyService } from "../services/companies.js";
 import { readBuiltInAgentMarker } from "../services/built-in-agent-metadata.js";
-import { reconcileBuiltInAgentsOnStartup } from "../services/built-in-agents.js";
+import { builtInAgentService, reconcileBuiltInAgentsOnStartup } from "../services/built-in-agents.js";
 
 const embeddedPostgresSupport = await getEmbeddedPostgresTestSupport();
 const describeEmbeddedPostgres = embeddedPostgresSupport.supported ? describe : describe.skip;
@@ -81,19 +81,28 @@ describeEmbeddedPostgres("companyService", () => {
     expect(rows.map((row) => row.issuePrefix).sort()).toEqual(["ARO", "AROA"]);
   });
 
-  it("auto-provisions one paused Reflection Coach bundle for a freshly created company", async () => {
+  it("does not auto-provision bundled built-in agents for a freshly created company", async () => {
     const created = await companyService(db).create({
       name: "Fresh Company",
     });
 
+    // A new company starts clean: the Reflection Coach and Summarizer are
+    // opt-in, not seeded by default for a new user.
     const agentRows = await db.select().from(agents).where(eq(agents.companyId, created.id));
-    const reflectionRows = agentRows.filter((row) => readBuiltInAgentMarker(row.metadata)?.key === "reflection-coach");
-    expect(reflectionRows).toHaveLength(1);
-    expect(reflectionRows[0]).toMatchObject({
+    expect(agentRows.filter((row) => readBuiltInAgentMarker(row.metadata))).toHaveLength(0);
+
+    // Startup reconcile leaves a fresh company untouched — nothing is created.
+    await reconcileBuiltInAgentsOnStartup(db);
+    const afterReconcileRows = await db.select().from(agents).where(eq(agents.companyId, created.id));
+    expect(afterReconcileRows.filter((row) => readBuiltInAgentMarker(row.metadata))).toHaveLength(0);
+
+    // The Reflection Coach remains available to enable on demand, and enabling
+    // it materializes its bundled skill + paused routine.
+    const enabled = await builtInAgentService(db).ensure(created.id, "reflection-coach");
+    expect(enabled.agent).toMatchObject({
       name: "Reflection Coach",
       status: "paused",
       budgetMonthlyCents: 0,
-      spentMonthlyCents: 0,
     });
 
     const [skill] = await db
@@ -110,10 +119,10 @@ describeEmbeddedPostgres("companyService", () => {
     const [routine] = await db
       .select()
       .from(routines)
-      .where(and(eq(routines.companyId, created.id), eq(routines.assigneeAgentId, reflectionRows[0]!.id)));
+      .where(and(eq(routines.companyId, created.id), eq(routines.assigneeAgentId, enabled.agentId!)));
     expect(routine).toMatchObject({
       status: "paused",
-      assigneeAgentId: reflectionRows[0]!.id,
+      assigneeAgentId: enabled.agentId,
       originKind: "built_in_agent_bundle",
       originId: "reflection-coach:recent-agent-reflection",
     });
@@ -122,10 +131,6 @@ describeEmbeddedPostgres("companyService", () => {
       kind: "schedule",
       enabled: false,
     });
-
-    await reconcileBuiltInAgentsOnStartup(db);
-    const afterReconcileRows = await db.select().from(agents).where(eq(agents.companyId, created.id));
-    expect(afterReconcileRows.filter((row) => readBuiltInAgentMarker(row.metadata)?.key === "reflection-coach")).toHaveLength(1);
   });
 
   it("archives companies by pausing runnable agents and cancelling active runs", async () => {

server/src/__tests__/instance-settings-service.test.ts

@@ -29,7 +29,6 @@ describe("instance settings service", () => {
       enableStreamlinedLeftNavigation: true,
       enableApps: false,
       enableConferenceRoomChat: false,
-      enableTaskChatRedesign: false,
       enableExternalObjects: false,
       enableSmokeLab: false,
       enablePipelines: false,