diff --git a/packages/cli/plugins/sentry-cli/skills/sentry-cli/references/init.md b/packages/cli/plugins/sentry-cli/skills/sentry-cli/references/init.md
index 9edc15c52..92d989351 100644
--- a/packages/cli/plugins/sentry-cli/skills/sentry-cli/references/init.md
+++ b/packages/cli/plugins/sentry-cli/skills/sentry-cli/references/init.md
@@ -18,7 +18,7 @@ Initialize Sentry in your project (experimental)
 **Flags:**
 - `-y, --yes - Accept non-interactive defaults (requires --features outside a TTY)`
 - `-n, --dry-run - Show what would happen without making changes`
-- `--features <value>... - Features to enable: errors,tracing,logs,replay,metrics,profiling,sourcemaps,crons,ai-monitoring`
+- `--features <value>... - Features to enable: errors,tracing,logs,replay,metrics,profiling,sourcemaps,crons,attachments,agent-tracing,mcp-observability`
 - `-t, --team <value> - Team slug to create the project under`
 - `--app <value> - App to initialize in a monorepo (required with --yes when multiple apps are detected)`
 - `--tui - Use the Ink-based interactive UI (default). Pass --no-tui to fall back to plain log output.`
diff --git a/packages/cli/src/commands/init.ts b/packages/cli/src/commands/init.ts
index 5501596fc..e5e5d582d 100644
--- a/packages/cli/src/commands/init.ts
+++ b/packages/cli/src/commands/init.ts
@@ -56,8 +56,13 @@ const FEATURE_ALIASES = {
   sourcemaps: "sourceMaps",
   sourceMaps: "sourceMaps",
   crons: "crons",
+  attachments: "attachments",
   "ai-monitoring": "aiMonitoring",
   aiMonitoring: "aiMonitoring",
+  "agent-tracing": "aiMonitoring",
+  agentTracing: "aiMonitoring",
+  "mcp-observability": "mcpObservability",
+  mcpObservability: "mcpObservability",
 } as const;
 
 const SUPPORTED_FEATURE_NAMES = [
@@ -69,7 +74,10 @@ const SUPPORTED_FEATURE_NAMES = [
   "profiling",
   "sourcemaps",
   "crons",
+  "attachments",
   "ai-monitoring",
+  "agent-tracing",
+  "mcp-observability",
 ] as const;
 
 const SUPPORTED_FEATURE_TEXT = SUPPORTED_FEATURE_NAMES.join(", ");
@@ -334,7 +342,7 @@ export const initCommand = buildCommand<
         kind: "parsed",
         parse: String,
         brief:
-          "Features to enable: errors,tracing,logs,replay,metrics,profiling,sourcemaps,crons,ai-monitoring",
+          "Features to enable: errors,tracing,logs,replay,metrics,profiling,sourcemaps,crons,attachments,agent-tracing,mcp-observability",
         variadic: true,
         optional: true,
       },
diff --git a/packages/cli/src/lib/init/clack-utils.ts b/packages/cli/src/lib/init/clack-utils.ts
index 8ba347fc2..25c103033 100644
--- a/packages/cli/src/lib/init/clack-utils.ts
+++ b/packages/cli/src/lib/init/clack-utils.ts
@@ -54,7 +54,7 @@ const FEATURE_INFO: Record<string, { label: string; description: string }> = {
       "Pinpoint the functions and lines of code responsible for performance issues",
   },
   logs: {
-    label: "Logging",
+    label: "Logs",
     description: "See logs in context with errors and performance issues",
   },
   metrics: {
@@ -68,11 +68,15 @@ const FEATURE_INFO: Record<string, { label: string; description: string }> = {
       "Turn minified production stack traces back into your original source code",
   },
   crons: {
-    label: "Crons",
+    label: "Crons & Uptime Monitors",
     description: "Detect failed, missed, or delayed scheduled jobs",
   },
+  attachments: {
+    label: "Attachments",
+    description: "Link user-supplied data to captured events",
+  },
   aiMonitoring: {
-    label: "AI Monitoring",
+    label: "Agent Tracing",
     description:
       "Understand AI calls, latency, token usage, cost, and failures",
   },
@@ -104,17 +108,18 @@ export function featureDescription(id: string): string | undefined {
 
 const FEATURE_DISPLAY_ORDER = [
   "errorMonitoring",
+  "logs",
   "sessionReplay",
   "performanceMonitoring",
-  "logs",
+  "aiMonitoring",
+  "attachments",
+  "crons",
   "metrics",
+  "mcpObservability",
   "profiling",
+  "reactFeatures",
   "sourceMaps",
-  "crons",
-  "aiMonitoring",
-  "mcpObservability",
   "userFeedback",
-  "reactFeatures",
 ];
 
 /** Sort features into the canonical order used by summaries and final output. */
@@ -134,7 +139,7 @@ export const STEP_LABELS: Record<string, string> = {
   "select-target-app": "Selecting target application",
   "resolve-dir": "Resolving project directory",
   "check-existing-sentry": "Checking for existing Sentry installation",
-  "detect-platform": "Detecting platform and framework",
+  "detect-platform": "Analyzing project and Sentry features",
   "ensure-sentry-project": "Setting up Sentry project",
   "select-features": "Selecting features",
   "plan-codemods": "Planning code modifications",
@@ -204,7 +209,7 @@ export const STEP_ACTIVE_LABELS: Record<string, string> = {
   "select-target-app": "Selecting target application...",
   "resolve-dir": "Resolving project directory...",
   "check-existing-sentry": "Checking for existing Sentry setup...",
-  "detect-platform": "Detecting framework and platform...",
+  "detect-platform": "Analyzing project and Sentry support...",
   "ensure-sentry-project": "Configuring Sentry project...",
   "select-features": "Preparing feature selection...",
   "plan-codemods": "Planning code changes...",
@@ -221,8 +226,8 @@ export const STEP_ACTIVE_LABELS: Record<string, string> = {
  * Falls back to the full label if a step isn't listed here.
  */
 export const STEP_LABELS_SHORT: Record<string, string> = {
-  "discover-context": "Analyzing project",
-  "detect-platform": "Detecting platform",
+  "discover-context": "Discovering project",
+  "detect-platform": "Checking Sentry support",
   "ensure-sentry-project": "Setting up project",
   "select-features": "Selecting features",
   "plan-codemods": "Planning changes",
@@ -256,9 +261,11 @@ export const STEP_PROGRESS_MESSAGES: Record<string, string[]> = {
   ],
   "detect-platform": [
     "Scanning project files...",
-    "Identifying framework and language...",
-    "Analyzing project configuration...",
-    "Determining SDK compatibility...",
+    "Identifying framework and runtime...",
+    "Matching the Sentry SDK...",
+    "Searching official Sentry docs...",
+    "Checking feature support...",
+    "Validating project recommendations...",
   ],
 };
 
diff --git a/packages/cli/src/lib/init/interactive.ts b/packages/cli/src/lib/init/interactive.ts
index acd8ff66b..eb98b930f 100644
--- a/packages/cli/src/lib/init/interactive.ts
+++ b/packages/cli/src/lib/init/interactive.ts
@@ -46,8 +46,12 @@ const DEFAULT_FEATURES = new Set<string>(DEFAULT_FEATURE_ORDER);
 const DEFAULT_FEATURE_RANK = new Map<string, number>(
   DEFAULT_FEATURE_ORDER.map((feature, index) => [feature, index])
 );
-// Feedback setup needs an in-app placement choice this wizard cannot make yet.
-const UNSUPPORTED_INIT_FEATURES = new Set(["userFeedback"]);
+// These features need code-level choices this selector cannot configure yet.
+const UNSUPPORTED_INIT_FEATURES = new Set([
+  "attachments",
+  "metrics",
+  "userFeedback",
+]);
 const FEATURE_SELECTION_CONTEXT =
   "Based on your project, these features are available to set up.";
 
@@ -69,7 +73,8 @@ function normalizeFeatureSelection(features: string[]): string[] {
   // the reviewed, resumed, and restored-on-Back selections identical.
   if (
     features.includes("aiMonitoring") ||
-    features.includes("mcpObservability")
+    features.includes("mcpObservability") ||
+    features.includes("profiling")
   ) {
     normalized.add("performanceMonitoring");
   }
diff --git a/packages/cli/src/lib/init/tools/agent-checkpoint.ts b/packages/cli/src/lib/init/tools/agent-checkpoint.ts
new file mode 100644
index 000000000..b97ddca4b
--- /dev/null
+++ b/packages/cli/src/lib/init/tools/agent-checkpoint.ts
@@ -0,0 +1,16 @@
+/**
+ * No-op init tool for acknowledging durable server-side agent checkpoints.
+ */
+
+import type { AgentCheckpointData, ToolResult } from "../types.js";
+import type { InitToolDefinition } from "./types.js";
+
+/** Resume a persisted agent session without performing local work. */
+export const agentCheckpointTool: InitToolDefinition<"agent-checkpoint"> = {
+  operation: "agent-checkpoint",
+  describe: (payload) => payload.detail ?? "Continuing repository analysis...",
+  execute: async (): Promise<ToolResult> => ({
+    data: { acknowledged: true } satisfies AgentCheckpointData,
+    ok: true,
+  }),
+};
diff --git a/packages/cli/src/lib/init/tools/registry.ts b/packages/cli/src/lib/init/tools/registry.ts
index e14feab56..340445a3d 100644
--- a/packages/cli/src/lib/init/tools/registry.ts
+++ b/packages/cli/src/lib/init/tools/registry.ts
@@ -1,5 +1,6 @@
 import { ApiError } from "../../errors.js";
 import type { ToolOperation, ToolPayload, ToolResult } from "../types.js";
+import { agentCheckpointTool } from "./agent-checkpoint.js";
 import { applyPatchsetTool } from "./apply-patchset.js";
 import {
   createSentryProjectTool,
@@ -16,6 +17,7 @@ import { formatToolError, validateToolSandbox } from "./shared.js";
 import type { AnyInitToolDefinition, ToolContext } from "./types.js";
 
 const toolDefinitions = [
+  agentCheckpointTool,
   listDirTool,
   readFilesTool,
   fileExistsBatchTool,
@@ -34,6 +36,7 @@ const toolRegistry = new Map<ToolOperation, AnyInitToolDefinition>(
 
 /** Sentry API operations never inspect or mutate the local filesystem. */
 const CWD_INDEPENDENT_OPERATIONS = new Set<ToolOperation>([
+  "agent-checkpoint",
   "create-sentry-project",
   "ensure-sentry-project",
 ]);
diff --git a/packages/cli/src/lib/init/types.ts b/packages/cli/src/lib/init/types.ts
index a62ce5d8a..965e09a88 100644
--- a/packages/cli/src/lib/init/types.ts
+++ b/packages/cli/src/lib/init/types.ts
@@ -76,6 +76,7 @@ export type InitProtocolEnvelope = {
 
 // Tool suspend payloads
 export type ToolPayload =
+  | AgentCheckpointPayload
   | ListDirPayload
   | ReadFilesPayload
   | FileExistsBatchPayload
@@ -89,6 +90,28 @@ export type ToolPayload =
 
 export type ToolOperation = ToolPayload["operation"];
 
+/**
+ * Suspend request used to continue a persisted server-side agent session.
+ * The CLI acknowledges it without inspecting or mutating the local project.
+ */
+export type AgentCheckpointPayload = {
+  /** Discriminator for locally handled workflow tools. */
+  type: "tool";
+  /** Stable no-op operation name. */
+  operation: "agent-checkpoint";
+  /** Optional progress copy displayed while the checkpoint is acknowledged. */
+  detail?: string;
+  /** Workflow filesystem root; not accessed by this operation. */
+  cwd: string;
+  /** Reserved operation arguments; currently always empty. */
+  params: Record<string, never>;
+};
+
+/** Explicit no-op acknowledgement validated by the init service boundary. */
+export type AgentCheckpointData = {
+  acknowledged: true;
+};
+
 export type ListDirPayload = {
   type: "tool";
   operation: "list-dir";
diff --git a/packages/cli/test/commands/init.test.ts b/packages/cli/test/commands/init.test.ts
index dfe699c28..583ecd15b 100644
--- a/packages/cli/test/commands/init.test.ts
+++ b/packages/cli/test/commands/init.test.ts
@@ -189,13 +189,18 @@ describe("init command func", () => {
       const ctx = makeContext();
       await func.call(ctx, {
         ...DEFAULT_FLAGS,
-        features: ["errors,tracing,replay,sourcemaps"],
+        features: [
+          "errors,tracing,replay,sourcemaps,attachments,agent-tracing,mcp-observability",
+        ],
       });
       expect(capturedArgs?.features).toEqual([
         "errorMonitoring",
         "performanceMonitoring",
         "sessionReplay",
         "sourceMaps",
+        "attachments",
+        "aiMonitoring",
+        "mcpObservability",
       ]);
     });
 
@@ -204,7 +209,7 @@ describe("init command func", () => {
       await func.call(ctx, {
         ...DEFAULT_FLAGS,
         features: [
-          "errorMonitoring,performanceMonitoring,sessionReplay,sourceMaps",
+          "errorMonitoring,performanceMonitoring,sessionReplay,sourceMaps,attachments,aiMonitoring,mcpObservability",
         ],
       });
       expect(capturedArgs?.features).toEqual([
@@ -212,6 +217,9 @@ describe("init command func", () => {
         "performanceMonitoring",
         "sessionReplay",
         "sourceMaps",
+        "attachments",
+        "aiMonitoring",
+        "mcpObservability",
       ]);
     });
 
@@ -223,7 +231,7 @@ describe("init command func", () => {
       });
       await expect(promise).rejects.toThrow(ValidationError);
       await expect(promise).rejects.toThrow(
-        "Supported features: errors, tracing, logs, replay, metrics, profiling, sourcemaps, crons, ai-monitoring"
+        "Supported features: errors, tracing, logs, replay, metrics, profiling, sourcemaps, crons, attachments, ai-monitoring, agent-tracing, mcp-observability"
       );
       expect(runWizardSpy).not.toHaveBeenCalled();
       expect(findProjectsSpy).not.toHaveBeenCalled();
diff --git a/packages/cli/test/lib/init/clack-utils.test.ts b/packages/cli/test/lib/init/clack-utils.test.ts
index 81a411009..82935cdaa 100644
--- a/packages/cli/test/lib/init/clack-utils.test.ts
+++ b/packages/cli/test/lib/init/clack-utils.test.ts
@@ -50,9 +50,11 @@ describe("featureLabel", () => {
   test("returns label for known feature", () => {
     expect(featureLabel("errorMonitoring")).toBe("Error Monitoring");
     expect(featureLabel("performanceMonitoring")).toBe("Tracing");
-    expect(featureLabel("logs")).toBe("Logging");
-    expect(featureLabel("crons")).toBe("Crons");
-    expect(featureLabel("aiMonitoring")).toBe("AI Monitoring");
+    expect(featureLabel("logs")).toBe("Logs");
+    expect(featureLabel("crons")).toBe("Crons & Uptime Monitors");
+    expect(featureLabel("attachments")).toBe("Attachments");
+    expect(featureLabel("aiMonitoring")).toBe("Agent Tracing");
+    expect(featureLabel("mcpObservability")).toBe("MCP Observability");
     expect(featureLabel("userFeedback")).toBe("User Feedback");
   });
 
@@ -87,6 +89,9 @@ describe("featureDescription", () => {
     expect(featureDescription("crons")).toBe(
       "Detect failed, missed, or delayed scheduled jobs"
     );
+    expect(featureDescription("attachments")).toBe(
+      "Link user-supplied data to captured events"
+    );
     expect(featureDescription("aiMonitoring")).toBe(
       "Understand AI calls, latency, token usage, cost, and failures"
     );
@@ -115,16 +120,18 @@ describe("sortFeatures", () => {
         "errorMonitoring",
         "sourceMaps",
         "crons",
+        "attachments",
         "aiMonitoring",
         "mcpObservability",
       ])
     ).toEqual([
       "errorMonitoring",
       "logs",
-      "sourceMaps",
-      "crons",
       "aiMonitoring",
+      "attachments",
+      "crons",
       "mcpObservability",
+      "sourceMaps",
       "userFeedback",
     ]);
   });
diff --git a/packages/cli/test/lib/init/interactive.test.ts b/packages/cli/test/lib/init/interactive.test.ts
index b3fc40f06..108b810c3 100644
--- a/packages/cli/test/lib/init/interactive.test.ts
+++ b/packages/cli/test/lib/init/interactive.test.ts
@@ -490,7 +490,7 @@ describe("handleMultiSelect", () => {
     });
   });
 
-  test("shows defaults first, sorts optional features, and omits User Feedback", async () => {
+  test("shows defaults first, sorts optional features, and omits unsupported features", async () => {
     const { ui, calls, respond } = createMockUI();
     respond.multiselect(["sessionReplay"]);
     respond.select("continue");
@@ -509,7 +509,9 @@ describe("handleMultiSelect", () => {
           "sessionReplay",
           "logs",
           "crons",
+          "attachments",
           "aiMonitoring",
+          "mcpObservability",
           "userFeedback",
         ],
       },
@@ -528,8 +530,8 @@ describe("handleMultiSelect", () => {
       "sessionReplay",
       "performanceMonitoring",
       "aiMonitoring",
-      "metrics",
       "crons",
+      "mcpObservability",
       "profiling",
       "sourceMaps",
     ]);
@@ -544,6 +546,8 @@ describe("handleMultiSelect", () => {
         text: "Based on your project, these features are available to set up.",
       },
     ]);
+    expect(multiselectCall?.options).not.toContain("metrics");
+    expect(multiselectCall?.options).not.toContain("attachments");
     expect(multiselectCall?.options).not.toContain("userFeedback");
 
     const reviewCall = calls.find((call) => call.kind === "select");
@@ -607,9 +611,10 @@ describe("handleMultiSelect", () => {
   });
 
   test.each([
-    "aiMonitoring",
-    "mcpObservability",
-  ])("review includes tracing when %s enables it implicitly", async (dependencyFeature) => {
+    ["aiMonitoring", "Agent Tracing"],
+    ["mcpObservability", "MCP Observability"],
+    ["profiling", "Profiling"],
+  ])("review includes tracing when %s enables it implicitly", async (dependencyFeature, dependencyLabel) => {
     const { ui, calls, respond } = createMockUI();
     respond.multiselect([dependencyFeature]);
     respond.select("continue");
@@ -636,9 +641,7 @@ describe("handleMultiSelect", () => {
     const reviewDetails = reviewCall?.details?.map((detail) => detail.text);
     expect(reviewDetails).toContain("✓ Error Monitoring");
     expect(reviewDetails).toContain("✓ Tracing");
-    expect(reviewDetails).toContain(
-      `✓ ${dependencyFeature === "aiMonitoring" ? "AI Monitoring" : "MCP Observability"}`
-    );
+    expect(reviewDetails).toContain(`✓ ${dependencyLabel}`);
   });
 
   test("Back restores the normalized AI selection including Tracing", async () => {
diff --git a/packages/cli/test/lib/init/tools/registry.test.ts b/packages/cli/test/lib/init/tools/registry.test.ts
index 7d89f7aa2..1d8689a43 100644
--- a/packages/cli/test/lib/init/tools/registry.test.ts
+++ b/packages/cli/test/lib/init/tools/registry.test.ts
@@ -19,6 +19,24 @@ function makeContext(): ResolvedInitContext {
 }
 
 describe("tool registry", () => {
+  test("acknowledges an agent checkpoint without local work", async () => {
+    const payload: ToolPayload = {
+      cwd: "/outside/project",
+      detail: "Reviewing official Sentry feature support...",
+      operation: "agent-checkpoint",
+      params: {},
+      type: "tool",
+    };
+
+    expect(describeTool(payload)).toBe(
+      "Reviewing official Sentry feature support..."
+    );
+    await expect(executeTool(payload, makeContext())).resolves.toEqual({
+      data: { acknowledged: true },
+      ok: true,
+    });
+  });
+
   test("describes tool payloads via the registered definition", () => {
     const payload: ToolPayload = {
       type: "tool",
diff --git a/packages/cli/test/lib/init/ui/ink-app.snapshot.test.tsx b/packages/cli/test/lib/init/ui/ink-app.snapshot.test.tsx
index 31a24e469..0b3d45556 100644
--- a/packages/cli/test/lib/init/ui/ink-app.snapshot.test.tsx
+++ b/packages/cli/test/lib/init/ui/ink-app.snapshot.test.tsx
@@ -275,10 +275,10 @@ describe("Ink App snapshot", () => {
     vi.useFakeTimers();
     try {
       const initialFrame = await renderActiveTaskFrameAfter(1);
-      expect(initialFrame).toContain("▶  Detecting platform");
+      expect(initialFrame).toContain("▶  Checking Sentry support");
 
       const pulsedFrame = await renderActiveTaskFrameAfter(601);
-      expect(pulsedFrame).toContain("▷  Detecting platform");
+      expect(pulsedFrame).toContain("▷  Checking Sentry support");
     } finally {
       vi.useRealTimers();
     }
diff --git a/packages/cli/test/lib/init/wizard-runner.test.ts b/packages/cli/test/lib/init/wizard-runner.test.ts
index e1a8e2869..9d34f52c4 100644
--- a/packages/cli/test/lib/init/wizard-runner.test.ts
+++ b/packages/cli/test/lib/init/wizard-runner.test.ts
@@ -1384,6 +1384,47 @@ describe("runWizard — resumeWithRetry stale-step recovery", () => {
     });
   });
 
+  test("sends the explicit agent-checkpoint acknowledgement", async () => {
+    const checkpointPayload: SuspendPayload = {
+      cwd: "/tmp/test",
+      detail: "Checking Sentry support for the detected project",
+      operation: "agent-checkpoint",
+      params: {},
+      protocolVersion: 1,
+      requestId: "5f61cbd5-1051-4b52-928f-06eb78ba40ee",
+      type: "tool",
+    };
+    mockStartResult = {
+      status: "suspended",
+      suspended: [["detect-platform"]],
+      steps: {
+        "detect-platform": { suspendPayload: checkpointPayload },
+      },
+    };
+    executeToolSpy.mockResolvedValue({
+      data: { acknowledged: true },
+      ok: true,
+    });
+    let capturedResume: Record<string, unknown> | undefined;
+    makeStaleStepRun((args) => {
+      capturedResume = args.resumeData as Record<string, unknown>;
+      return Promise.resolve({ status: "success", result: { exitCode: 0 } });
+    });
+
+    await runWizard(makeOptions());
+
+    expect(executeToolSpy).toHaveBeenCalledWith(
+      expect.objectContaining({ operation: "agent-checkpoint" }),
+      expect.anything()
+    );
+    expect(capturedResume).toMatchObject({
+      data: { acknowledged: true },
+      ok: true,
+      protocolVersion: 1,
+      requestId: checkpointPayload.requestId,
+    });
+  });
+
   test("uses request identity instead of mutable payload details during recovery", async () => {
     vi.useFakeTimers();
     const requestId = "8c7ee6b9-e955-4514-9164-f01844584a28";