diff --git a/scripts/bench-cli-startup.ts b/scripts/bench-cli-startup.ts
index cebcd9dc775..cbec863e38d 100644
--- a/scripts/bench-cli-startup.ts
+++ b/scripts/bench-cli-startup.ts
@@ -449,7 +449,7 @@ function parseFlagValue(flag: string): string | undefined {
     return undefined;
   }
   const value = process.argv[idx + 1];
-  if (!value || value.startsWith("--")) {
+  if (!value || value.startsWith("-")) {
     throw new Error(`${flag} requires a value`);
   }
   return value;
@@ -462,7 +462,8 @@ function hasFlag(flag: string): boolean {
 function parseRepeatableFlag(flag: string): string[] {
   const values: string[] = [];
   for (let i = 0; i < process.argv.length; i += 1) {
-    if (process.argv[i] === flag && process.argv[i + 1]) {
+    const value = process.argv[i + 1];
+    if (process.argv[i] === flag && value && !value.startsWith("-")) {
       values.push(process.argv[i + 1]);
     }
   }
@@ -474,7 +475,7 @@ function validateCliArgs(argv: readonly string[] = process.argv.slice(2)): void
     const arg = argv[index];
     if (VALUE_FLAGS.has(arg)) {
       const value = argv[index + 1];
-      if (!value || value.startsWith("--")) {
+      if (!value || value.startsWith("-")) {
         throw new Error(`${arg} requires a value`);
       }
       index += 1;
@@ -1161,12 +1162,12 @@ function readBenchmarkComparisonForTesting(
 }
 
 async function main(): Promise<void> {
+  validateCliArgs();
   if (hasFlag("--help")) {
     printUsage();
     return;
   }
 
-  validateCliArgs();
   const options = parseOptions();
   if (options.compareBaseline || options.compareCandidate) {
     if (!options.compareBaseline || !options.compareCandidate) {
diff --git a/scripts/bench-model.ts b/scripts/bench-model.ts
index 4424b6aab37..69b9b050763 100644
--- a/scripts/bench-model.ts
+++ b/scripts/bench-model.ts
@@ -33,7 +33,7 @@ class CliArgumentError extends Error {
 
 function readValue(argv: string[], index: number, flag: string): string {
   const value = argv[index + 1]?.trim() ?? "";
-  if (!value || value.startsWith("--")) {
+  if (!value || value.startsWith("-")) {
     throw new CliArgumentError(`${flag} requires a value`);
   }
   return value;
diff --git a/scripts/qa/ux-matrix-evidence-producer.ts b/scripts/qa/ux-matrix-evidence-producer.ts
index 01d3ed0a2fb..1bc7ba84230 100644
--- a/scripts/qa/ux-matrix-evidence-producer.ts
+++ b/scripts/qa/ux-matrix-evidence-producer.ts
@@ -53,12 +53,16 @@ Options:
 
 function readOptionValue(argv: readonly string[], index: number, arg: string) {
   const value = argv[index + 1] ?? "";
-  if (!value || value.startsWith("--")) {
+  if (!value || value.startsWith("-")) {
     throw new Error(`${arg} requires a value`);
   }
   return value;
 }
 
+function isHelpRequest(argv: readonly string[]) {
+  return argv.length === 1 && (argv[0] === "--help" || argv[0] === "-h");
+}
+
 function parseOptions(argv: readonly string[]): ProducerOptions {
   let artifactBase = "";
   let repoRoot = process.cwd();
@@ -698,7 +702,7 @@ export async function runUxMatrixEvidenceProducer(options: ProducerOptions) {
 if (import.meta.url === pathToFileURL(process.argv[1] ?? "").href) {
   (async () => {
     const cliArgs = process.argv.slice(2);
-    if (cliArgs.includes("--help") || cliArgs.includes("-h")) {
+    if (isHelpRequest(cliArgs)) {
       console.log(usage());
       return;
     }
diff --git a/test/scripts/bench-cli-startup.test.ts b/test/scripts/bench-cli-startup.test.ts
index 5b02033a23c..0db3312edd3 100644
--- a/test/scripts/bench-cli-startup.test.ts
+++ b/test/scripts/bench-cli-startup.test.ts
@@ -36,6 +36,26 @@ describe("bench-cli-startup", () => {
     expect(result.stderr).not.toContain("\n    at ");
   });
 
+  it("rejects short flag values before running benchmarks", () => {
+    expect(() => testing.validateCliArgs(["--output", "-h"])).toThrow("--output requires a value");
+    expect(() => testing.validateCliArgs(["--case", "-h"])).toThrow("--case requires a value");
+
+    const result = spawnSync(
+      process.execPath,
+      ["--import", "tsx", "scripts/bench-cli-startup.ts", "--output", "-h"],
+      {
+        cwd: join(__dirname, "../.."),
+        encoding: "utf8",
+      },
+    );
+
+    expect(result.status).toBe(1);
+    expect(result.stdout).toBe("");
+    expect(result.stderr.trim()).toBe("--output requires a value");
+    expect(result.stderr).not.toContain("Node.js");
+    expect(result.stderr).not.toContain("\n    at ");
+  });
+
   it.runIf(process.platform !== "win32")(
     "cleans timed-out benchmark process groups when the leader exits first",
     () => {
@@ -43,7 +63,7 @@ describe("bench-cli-startup", () => {
       const tmpDir = tempDirs.make("openclaw-cli-startup-timeout-group-");
       const entryPath = join(tmpDir, "entry.mjs");
       const childPidPath = join(tmpDir, "child.pid");
-      let childPid = 0;
+      let childPid: number | undefined;
       try {
         writeFileSync(
           entryPath,
@@ -93,7 +113,7 @@ describe("bench-cli-startup", () => {
         expect(result.stderr).toContain("version sample 1: timed out");
         expect(isProcessAlive(childPid)).toBe(false);
       } finally {
-        if (childPid && isProcessAlive(childPid)) {
+        if (childPid !== undefined && isProcessAlive(childPid)) {
           process.kill(childPid, "SIGKILL");
         }
         tempDirs.cleanup();
diff --git a/test/scripts/bench-model.test.ts b/test/scripts/bench-model.test.ts
index 51f31a33437..12fe5977011 100644
--- a/test/scripts/bench-model.test.ts
+++ b/test/scripts/bench-model.test.ts
@@ -47,6 +47,18 @@ describe("scripts/bench-model", () => {
     expect(result.stderr).not.toContain("Missing ANTHROPIC_API_KEY");
   });
 
+  it("rejects short flag values before checking provider credentials", () => {
+    expect(() => testing.parseArgs(["--prompt", "-h"])).toThrow("--prompt requires a value");
+    expect(() => testing.parseArgs(["--runs", "-h"])).toThrow("--runs requires a value");
+
+    const result = runBenchModel(["--prompt", "-h"]);
+
+    expect(result.status).toBe(1);
+    expect(result.stdout).toBe("");
+    expect(result.stderr.trim()).toBe("--prompt requires a value");
+    expect(result.stderr).not.toContain("Missing ANTHROPIC_API_KEY");
+  });
+
   it("prints help without checking provider credentials", () => {
     const result = runBenchModel(["--help"]);
 
diff --git a/test/scripts/qa-ux-matrix-evidence-producer.test.ts b/test/scripts/qa-ux-matrix-evidence-producer.test.ts
index c6be91220cb..1bac3f72309 100644
--- a/test/scripts/qa-ux-matrix-evidence-producer.test.ts
+++ b/test/scripts/qa-ux-matrix-evidence-producer.test.ts
@@ -52,6 +52,20 @@ describe("QA UX Matrix evidence producer CLI", () => {
     expectNoNodeStack(result.stderr);
   });
 
+  it("reports short flag values without treating them as help", () => {
+    const artifactBaseResult = runCli("--artifact-base", "-h");
+    const repoRootResult = runCli("--artifact-base", "/tmp/openclaw-ux-test", "--repo-root", "-h");
+
+    expect(artifactBaseResult.status).toBe(1);
+    expect(artifactBaseResult.stdout).toBe("");
+    expect(artifactBaseResult.stderr.trim()).toBe("--artifact-base requires a value");
+    expectNoNodeStack(artifactBaseResult.stderr);
+    expect(repoRootResult.status).toBe(1);
+    expect(repoRootResult.stdout).toBe("");
+    expect(repoRootResult.stderr.trim()).toBe("--repo-root requires a value");
+    expectNoNodeStack(repoRootResult.stderr);
+  });
+
   it("sanitizes local checkout paths from generated evidence artifacts", () => {
     const artifactBase = fs.mkdtempSync(path.join(os.tmpdir(), "openclaw-ux-evidence-test-"));
     const fakeRepoRoot = fs.mkdtempSync(path.join(os.tmpdir(), "openclaw-ux-repo-test-"));