Skip to content

armgit / rea

Public Imported from GitHub
0

Reverse engineer anything with agents, from app behavior down to native binaries.

fix(cli): preserve output and comparison input validation (#981)

morlutod5c43d65f0f6
diff --git a/src/cli/browserAdvancedCommands.ts b/src/cli/browserAdvancedCommands.ts
index 2d299503f..0fd49960e 100644
--- a/src/cli/browserAdvancedCommands.ts
+++ b/src/cli/browserAdvancedCommands.ts
@@ -10,7 +10,10 @@ import {
} from "../application/BrowserObservationService.js";
import { CdpBrowserProvider } from "../browser/CdpBrowserProvider.js";
import { logCliCommand } from "../cliLogging.js";
-import { AnalysisInputError } from "../domain/analysisErrorCore.js";
+import {
+ AnalysisInputError,
+ type AnalysisInputIssue,
+} from "../domain/analysisErrorCore.js";
import { browserCaptureComparisonInputSchema } from "../domain/browserCaptureComparison.js";
import { discoverWebMcpToolsInputSchema } from "../domain/webMcpDiscovery.js";
import {
@@ -89,19 +92,32 @@ const registerCaptureDiff = (
}),
run: ({ args, options }) =>
logCliCommand(logger, "compare-web-captures", async () => {
- const before = parseJson(args.beforeJson);
- const after = parseJson(args.afterJson);
+ const before = safeParseJson(args.beforeJson);
+ const after = safeParseJson(args.afterJson);
+ const normalization = safeParseJson(options.normalizationJson);
+ if (!before.ok)
+ return inputError("compare_web_captures", [
+ invalidJsonIssue("before"),
+ ]);
+ if (!after.ok)
+ return inputError("compare_web_captures", [
+ invalidJsonIssue("after"),
+ ]);
+ if (!normalization.ok)
+ return inputError("compare_web_captures", [
+ invalidJsonIssue("normalization"),
+ ]);
const scenarioComparison =
browserCaptureComparisonInputSchema.safeParse({
- before_scenario: before,
- after_scenario: after,
- normalization: parseJson(options.normalizationJson),
+ before_scenario: before.value,
+ after_scenario: after.value,
+ normalization: normalization.value,
});
const parsed = scenarioComparison.success
? scenarioComparison
: browserCaptureComparisonInputSchema.safeParse({
- before,
- after,
+ before: before.value,
+ after: after.value,
});
if (!parsed.success) return inputError("compare_web_captures");
const result = await compareWebCaptureEvidence(
@@ -186,5 +202,14 @@ const parseJson = (value: string): unknown => {
return parsed.ok ? parsed.value : undefined;
};
-const inputError = (operation: string): JsonValue =>
- browserCliError(new AnalysisInputError(operation));
+const invalidJsonIssue = (field: string): AnalysisInputIssue => ({
+ path: [field],
+ reason: "invalid_format",
+ expected: "JSON",
+});
+
+const inputError = (
+ operation: string,
+ issues: readonly AnalysisInputIssue[] = [],
+): JsonValue =>
+ browserCliError(new AnalysisInputError(operation, undefined, issues));
diff --git a/src/cliOutput.ts b/src/cliOutput.ts
index 96f65a7e7..063a6d452 100644
--- a/src/cliOutput.ts
+++ b/src/cliOutput.ts
@@ -6,6 +6,7 @@ const UNSUPPORTED_OUTPUT_COMBINATION_MESSAGE =
"Token windows cannot preserve structured output. Remove --token-limit/--token-offset and use --filter-output or command pagination.";
type StructuredOutputFormat = "json" | "jsonl" | "yaml";
+const INCUR_OUTPUT_FORMATS = new Set(["toon", "json", "yaml", "md", "jsonl"]);
export type CliOutputArgumentValidation =
| { readonly ok: true }
@@ -21,6 +22,7 @@ type JsonRecord = Record<string, unknown>;
interface ParsedCliOutputArguments {
readonly filterOutput: boolean;
readonly format: string;
+ readonly parseError: boolean;
readonly tokenWindow: boolean;
}
@@ -50,40 +52,45 @@ const parseCliOutputArguments = (
}
if (argument === "--format") {
const value = arguments_[index + 1];
- if (value !== undefined) {
+ if (value) {
+ // Incur stops parsing at an invalid format before applying any later
+ // output controls. Let it report that malformed flag itself.
+ if (!INCUR_OUTPUT_FORMATS.has(value))
+ return { filterOutput, format, parseError: true, tokenWindow: false };
format = value;
index += 1;
}
continue;
}
if (argument === "--filter-output") {
- if (arguments_[index + 1] !== undefined) {
+ if (arguments_[index + 1]) {
filterOutput = true;
index += 1;
}
continue;
}
if (argument === "--token-limit" || argument === "--token-offset") {
- if (arguments_[index + 1] !== undefined) {
+ const value = arguments_[index + 1];
+ if (value) {
+ const numericValue = Number(value);
+ if (!Number.isFinite(numericValue) || value.trim() === "")
+ return { filterOutput, format, parseError: true, tokenWindow: false };
tokenWindow = true;
index += 1;
}
continue;
}
- if (
- argument?.startsWith("--token-limit=") ||
- argument?.startsWith("--token-offset=")
- )
- tokenWindow = true;
}
- return { filterOutput, format, tokenWindow };
+ return { filterOutput, format, parseError: false, tokenWindow };
};
/** Reject text-window controls that would corrupt a structured document. */
export const validateCliOutputArguments = (
arguments_: readonly string[],
): CliOutputArgumentValidation => {
- const { format, tokenWindow } = parseCliOutputArguments(arguments_);
+ const { format, parseError, tokenWindow } =
+ parseCliOutputArguments(arguments_);
+ if (parseError) return { ok: true };
if (
tokenWindow &&
(format === "json" || format === "jsonl" || format === "yaml")
@@ -101,7 +108,9 @@ export const validateCliOutputArguments = (
export const renderEmptyFilteredCliOutput = (
arguments_: readonly string[],
): string | undefined => {
- const { filterOutput, format } = parseCliOutputArguments(arguments_);
+ const { filterOutput, format, parseError } =
+ parseCliOutputArguments(arguments_);
+ if (parseError) return undefined;
if (
filterOutput &&
(format === "json" || format === "jsonl" || format === "yaml")
diff --git a/tests/boundary/cli/browserCli.test.ts b/tests/boundary/cli/browserCli.test.ts
index 6aab7fe3e..9ddd0ac9c 100644
--- a/tests/boundary/cli/browserCli.test.ts
+++ b/tests/boundary/cli/browserCli.test.ts
@@ -108,6 +108,57 @@ describe("browser CLI parity", () => {
alignment: { status: "aligned", aligned_steps: 2 },
},
});
+ const intentionalPolicy = await runCli(
+ [
+ "compare-web-captures",
+ JSON.stringify(scenarioCapture),
+ JSON.stringify(scenarioCapture),
+ "--normalization-json",
+ JSON.stringify({
+ rules: [
+ {
+ rule_id: "domain",
+ artifacts: ["url"],
+ match: "app.example.test",
+ replacement: "normalized.test",
+ },
+ ],
+ }),
+ "--json",
+ ],
+ process.env,
+ );
+ expect(intentionalPolicy).toMatchObject({
+ normalized_result: {
+ normalization: { rules: [{ rule_id: "domain" }] },
+ },
+ });
+
+ const malformedPolicy = await runCli(
+ [
+ "compare-web-captures",
+ JSON.stringify(scenarioCapture),
+ JSON.stringify(scenarioCapture),
+ "--normalization-json",
+ "{invalid-policy",
+ "--json",
+ ],
+ process.env,
+ );
+ expect(malformedPolicy).toMatchObject({
+ error: "Browser observation failed",
+ code: "invalid_request",
+ details: {
+ operation: "compare_web_captures",
+ issues: [
+ {
+ path: ["normalization"],
+ reason: "invalid_format",
+ expected: "JSON",
+ },
+ ],
+ },
+ });
},
INTEGRATION_TEST_TIMEOUT_MS,
);
diff --git a/tests/boundary/cli/cliOutput.test.ts b/tests/boundary/cli/cliOutput.test.ts
index 22fa1b0a6..c31ebf47a 100644
--- a/tests/boundary/cli/cliOutput.test.ts
+++ b/tests/boundary/cli/cliOutput.test.ts
@@ -57,7 +57,9 @@ describe("CLI output argument and sanitization boundary", () => {
validateCliOutputArguments(["providers", "--token-count", "--json"]),
).toEqual({ ok: true });
});
+});
+describe("CLI output valued-flag parsing", () => {
workspaceCliTest(
"preserves complete JSON errors when native builtin flags follow unusual arguments",
async ({ cli }) => {
@@ -92,6 +94,70 @@ describe("CLI output argument and sanitization boundary", () => {
CLI_INTEGRATION_TIMEOUT_MS,
);
+ workspaceCliTest(
+ "does not treat empty or equals-style output values as active builtins",
+ async ({ cli }) => {
+ for (const arguments_ of [
+ ["capabilities", "--json", "--format", ""],
+ ["capabilities", "--json", "--filter-output", ""],
+ ["capabilities", "--json", "--token-limit", ""],
+ ["capabilities", "--json", "--token-limit=5"],
+ ]) {
+ const result = await cli.run({ arguments: arguments_ });
+ expect(result.exitCode).toBe(1);
+ expect(result.json).toBeDefined();
+ expect(result.stdout).not.toContain("[truncated:");
+ expect(JSON.stringify(result.json)).not.toContain(
+ "UNSUPPORTED_OUTPUT_COMBINATION",
+ );
+ }
+ const validTokenWindow = await cli.run({
+ arguments: ["capabilities", "--json", "--token-limit", "5"],
+ });
+ expect(validTokenWindow.exitCode).toBe(1);
+ expect(validTokenWindow.json).toMatchObject({
+ ok: false,
+ error: { code: "UNSUPPORTED_OUTPUT_COMBINATION" },
+ });
+ const emptyFormatWithTokenWindow = await cli.run({
+ arguments: [
+ "capabilities",
+ "--json",
+ "--format",
+ "",
+ "--token-limit",
+ "5",
+ ],
+ });
+ expect(emptyFormatWithTokenWindow.exitCode).toBe(1);
+ expect(emptyFormatWithTokenWindow.stdout).not.toContain("[truncated:");
+ expect(emptyFormatWithTokenWindow.json).toMatchObject({
+ ok: false,
+ error: { code: "UNSUPPORTED_OUTPUT_COMBINATION" },
+ });
+
+ for (const arguments_ of [
+ ["capabilities", "--json", "--token-limit", "not-a-number"],
+ [
+ "capabilities",
+ "--format",
+ "unsupported-format",
+ "--json",
+ "--token-limit",
+ "5",
+ ],
+ ]) {
+ const result = await cli.run({ arguments: arguments_ });
+ expect(result.exitCode).toBe(1);
+ expect(result.stdout).not.toContain("UNSUPPORTED_OUTPUT_COMBINATION");
+ expect(result.stdout).not.toContain("[truncated:");
+ }
+ },
+ CLI_INTEGRATION_TIMEOUT_MS,
+ );
+});
+
+describe("CLI output argument and sanitization boundary", () => {
it("preserves normal output and sanitizes text and JSON validation errors", () => {
expect(sanitizeCliOutput("result: ok\n")).toBe("result: ok\n");
expect(sanitizeCliOutput("result: VALIDATION_ERROR\n")).toBe(