diff --git a/docker/entrypoint.sh b/docker/entrypoint.sh
index 347ba21c..a8ca4894 100644
--- a/docker/entrypoint.sh
+++ b/docker/entrypoint.sh
@@ -37,7 +37,7 @@ for argument do
             fi
             ;;
         --output-dir|--workers|--mode|--model|--effort|--provider|\
-            --knowledge-base|--max-attempts|--plugin-path|--python|\
+            --knowledge-base|--max-attempts|--max-cost|--plugin-path|--python|\
             --codex|--filter-output|--format|\
             --scan-prompt-file|--post-scan-prompt-file|\
             --token-limit|--token-offset)
diff --git a/sdk/typescript/README.md b/sdk/typescript/README.md
index e32c4e8d..0a08489b 100644
--- a/sdk/typescript/README.md
+++ b/sdk/typescript/README.md
@@ -217,7 +217,7 @@ npx @openai/codex-security scan /path/to/repository --mode deep --workers 2 --su
 npx @openai/codex-security install-hook
 npx @openai/codex-security bulk-scan
 npx @openai/codex-security bulk-scan --model gpt-5.6-terra --effort high
-npx @openai/codex-security bulk-scan --workers 4 --mode deep --max-attempts 3
+npx @openai/codex-security bulk-scan --workers 4 --mode deep --max-attempts 3 --max-cost 25
 npx @openai/codex-security bulk-scan repositories.csv --output-dir /path/outside/repositories/security-scans --workers 4 --knowledge-base /path/to/threat-models --knowledge-base /path/to/architecture.pdf
 npx @openai/codex-security bulk-scan repositories.csv --output-dir /path/outside/repositories/security-scans --scan-prompt-file scan.md --post-scan-prompt-file follow-up.md
 npx @openai/codex-security scans list /path/to/repository
@@ -486,6 +486,7 @@ it returns a sealed partial report with any completed findings and lists
 unvalidated candidates as follow-up work. Requests already in progress can
 finish above the limit; preparing the partial report makes no additional model
 requests. Incomplete coverage retains its existing exit code.
+For `bulk-scan`, the limit applies separately to each repository attempt.
 
 Run `npx @openai/codex-security scan --help` or `npx @openai/codex-security bulk-scan --help`
 for the complete CLI references.
diff --git a/sdk/typescript/src/cli.ts b/sdk/typescript/src/cli.ts
index e8483157..8d89d99f 100644
--- a/sdk/typescript/src/cli.ts
+++ b/sdk/typescript/src/cli.ts
@@ -1455,6 +1455,13 @@ export async function main(
           .positive()
           .default(1)
           .describe("Maximum scan attempts per repository."),
+        maxCost: z
+          .number()
+          .positive()
+          .optional()
+          .describe(
+            "Stop each repository attempt if estimated USD cost exceeds AMOUNT.",
+          ),
         pluginPath: z
           .string()
           .min(1)
@@ -1535,6 +1542,9 @@ export async function main(
             workers: options.workers,
             mode: options.mode,
             maxAttempts: options.maxAttempts,
+            ...(options.maxCost === undefined
+              ? {}
+              : { maxCostUsd: options.maxCost }),
             knowledgeBasePaths: options.knowledgeBase,
             ...prompts,
             config: {
diff --git a/sdk/typescript/src/multiscan.ts b/sdk/typescript/src/multiscan.ts
index 2c96206c..fe151774 100644
--- a/sdk/typescript/src/multiscan.ts
+++ b/sdk/typescript/src/multiscan.ts
@@ -20,7 +20,7 @@ import Papa from "papaparse";
 import type { CodexSecurity } from "./api.js";
 import type { CodexSecurityConfig } from "./config.js";
 import type { ScanCost } from "./cost.js";
-import { safeErrorMessage } from "./errors.js";
+import { safeErrorMessage, ScanCostLimitExceededError } from "./errors.js";
 import type { CoverageDocument } from "./models.js";
 import type { ScanMode } from "./targets.js";
 import { resolveTrustedExecutable } from "./trusted-executable.js";
@@ -62,6 +62,7 @@ export interface MultiscanOptions {
   workers: number;
   mode: ScanMode;
   maxAttempts: number;
+  maxCostUsd?: number;
   scanPrompt?: string;
   postScanPrompt?: string;
   config: CodexSecurityConfig;
@@ -198,6 +199,7 @@ async function runCampaign(
         let warning: string | undefined;
         let coverage: CoverageDocument["completeness"] | undefined;
         let cost: Readonly<ScanCost> | null = null;
+        let exhaustedBudget = false;
         try {
           await mkdir(dirname(scanDir), { recursive: true, mode: 0o700 });
           await rm(checkout, { recursive: true, force: true });
@@ -233,6 +235,9 @@ async function runCampaign(
             ...(options.postScanPrompt === undefined
               ? {}
               : { postScanPrompt: options.postScanPrompt }),
+            ...(options.maxCostUsd === undefined
+              ? {}
+              : { maxCostUsd: options.maxCostUsd }),
             onWarning: (warning) =>
               options.onProgress?.({ ...progress, status: "started", warning }),
             ...(options.signal === undefined ? {} : { signal: options.signal }),
@@ -249,6 +254,10 @@ async function runCampaign(
           }
         } catch (error) {
           if (options.signal?.aborted === true) options.signal.throwIfAborted();
+          if (error instanceof ScanCostLimitExceededError) {
+            cost = error.cost;
+            exhaustedBudget = true;
+          }
           failure = safeErrorMessage(error);
         } finally {
           await rm(checkout, { recursive: true, force: true });
@@ -283,6 +292,10 @@ async function runCampaign(
           else incomplete += 1;
           break;
         }
+        if (exhaustedBudget) {
+          failed += 1;
+          break;
+        }
         if (retry === options.maxAttempts - 1) failed += 1;
       }
     }
@@ -521,7 +534,10 @@ async function recoverLock(
 async function ensureManifest(
   path: string,
   tasks: MultiscanTask[],
-  options: Pick<MultiscanOptions, "scanPrompt" | "postScanPrompt">,
+  options: Pick<
+    MultiscanOptions,
+    "scanPrompt" | "postScanPrompt" | "maxCostUsd"
+  >,
 ): Promise<void> {
   const expected = `${JSON.stringify(
     {
@@ -533,6 +549,9 @@ async function ensureManifest(
       ...(options.postScanPrompt === undefined
         ? {}
         : { postScanPrompt: options.postScanPrompt }),
+      ...(options.maxCostUsd === undefined
+        ? {}
+        : { maxCostUsd: options.maxCostUsd }),
     },
     null,
     2,
diff --git a/sdk/typescript/tests-ts/container-entrypoint.test.ts b/sdk/typescript/tests-ts/container-entrypoint.test.ts
index 796678a0..ff25a177 100644
--- a/sdk/typescript/tests-ts/container-entrypoint.test.ts
+++ b/sdk/typescript/tests-ts/container-entrypoint.test.ts
@@ -169,6 +169,7 @@ describe("customer container entrypoint", () => {
   testPosix("accepts CSVs after global and bulk-scan options", async () => {
     for (const arguments_ of [
       ["bulk-scan", "--workers", "2", "/input/repositories.csv"],
+      ["bulk-scan", "--max-cost", "12.50", "/input/repositories.csv"],
       ["bulk-scan", "bulk-scan", "--output-dir", "/output"],
       [
         "--format",
@@ -260,6 +261,7 @@ describe("customer container entrypoint", () => {
       for (const arguments_ of [
         ["bulk-scan"],
         ["bulk-scan", "--workers", "8"],
+        ["bulk-scan", "--max-cost", "12.50"],
         ["bulk-scan", "--scan-prompt-file", "prompt.md"],
         ["bulk-scan", "--post-scan-prompt-file", "post-prompt.md"],
         ["--format", "json", "bulk-scan", "--mode", "deep"],
diff --git a/sdk/typescript/tests-ts/multiscan.test.ts b/sdk/typescript/tests-ts/multiscan.test.ts
index c8244933..3dc00390 100644
--- a/sdk/typescript/tests-ts/multiscan.test.ts
+++ b/sdk/typescript/tests-ts/multiscan.test.ts
@@ -18,6 +18,7 @@ import { hostname, tmpdir } from "node:os";
 import { join } from "node:path";
 import { afterEach, describe, expect, spyOn, test } from "bun:test";
 import { main } from "../src/cli.js";
+import { ScanCostLimitExceededError } from "../src/errors.js";
 import type { ScanResult } from "../src/result.js";
 import { buildGitHubCredentialArgs, runMultiscan } from "../src/multiscan.js";
 import { resolveTrustedExecutable } from "../src/trusted-executable.js";
@@ -204,11 +205,13 @@ describe("multiscan", () => {
             "Review boundaries.\n\nFocus on authentication, authorization.",
           );
           expect(scanOptions.postScanPrompt).toBe("Draft confirmed fixes.");
+          expect(scanOptions.maxCostUsd).toBe(12.5);
           return await completedScan(scanOptions.outputDir!);
         }),
         {
           scanPrompt: "Review boundaries.",
           postScanPrompt: "Draft confirmed fixes.",
+          maxCostUsd: 12.5,
         },
       ),
     );
@@ -222,6 +225,7 @@ describe("multiscan", () => {
     ).toMatchObject({
       scanPrompt: "Review boundaries.",
       postScanPrompt: "Draft confirmed fixes.",
+      maxCostUsd: 12.5,
       tasks: [
         { id: "payments", prompt: "Focus on authentication, authorization." },
       ],
@@ -259,6 +263,89 @@ describe("multiscan", () => {
     ]);
   });
 
+  test("records an exhausted repository budget without retrying the scan", async () => {
+    const paths = await fixture();
+    const source = await repository(paths.root, "over-budget");
+    await writeFile(
+      paths.input,
+      `id,repository,revision\nover-budget,${source.path},${source.revision}\n`,
+    );
+    const cost = {
+      model: "gpt-5.6-sol",
+      inputTokens: 1_250,
+      cachedInputTokens: 200,
+      cacheWriteInputTokens: 0,
+      outputTokens: 30,
+      estimatedUsd: 25.25,
+    };
+    let attempts = 0;
+
+    const summary = await runMultiscan(
+      options(
+        paths,
+        client(async (_repository, scanOptions = {}) => {
+          attempts += 1;
+          throw new ScanCostLimitExceededError(
+            25,
+            cost,
+            scanOptions.outputDir!,
+          );
+        }),
+        { maxAttempts: 3, maxCostUsd: 25 },
+      ),
+    );
+
+    expect(attempts).toBe(1);
+    expect(summary).toMatchObject({ completed: 0, failed: 1 });
+    expect(await results(summary.resultsPath)).toMatchObject([
+      { id: "over-budget", status: "failed", attempt: 1, cost },
+    ]);
+  });
+
+  test("forwards a bulk CLI cost limit and rejects zero", async () => {
+    const paths = await fixture();
+    const source = await repository(paths.root, "sample");
+    await writeFile(
+      paths.input,
+      `id,repository,revision\nsample,${source.path},${source.revision}\n`,
+    );
+    const stdout = capture();
+    const stderr = capture();
+    let scanOptions: unknown;
+
+    expect(
+      await main(
+        [
+          "bulk-scan",
+          "repositories.csv",
+          "--output-dir",
+          "results",
+          "--max-cost",
+          "12.5",
+          "--json",
+        ],
+        stdout.stream,
+        stderr.stream,
+        dependencies({
+          currentDirectory: paths.root,
+          onTurn: (_repository, options) => (scanOptions = options),
+        }),
+      ),
+    ).toBe(0);
+    expect(scanOptions).toMatchObject({ maxCostUsd: 12.5 });
+
+    const invalid = capture();
+    expect(
+      await main(
+        ["bulk-scan", "--max-cost=0"],
+        capture().stream,
+        invalid.stream,
+        dependencies({ currentDirectory: paths.root }),
+      ),
+    ).toBe(2);
+    expect(invalid.text()).toContain("expected number to be >0");
+  });
+
   test("surfaces optional post-scan warnings without failing completed scans", async () => {
     const paths = await fixture();
     const source = await repository(paths.root, "follow-up-warning");
@@ -1374,6 +1461,7 @@ describe("multiscan", () => {
     for (const prompts of [
       { scanPrompt: "Review different boundaries." },
       { postScanPrompt: "Draft confirmed fixes." },
+      { maxCostUsd: 12.5 },
     ]) {
       await expect(
         runMultiscan(options(paths, security, prompts)),