Skip to content

OpenAI's Codex Security CLI and TypeScript SDK for finding, validating, and fixing security vulnerabilities. npm: https://www.npmjs.com/package/@openai/codex-security

fix(csv): preserve finding occurrences and literal apostrophes (#1247)

mldangelo-oaibcda23b2a1b9
diff --git a/plugins/codex-security/scripts/finalize_scan_contract.py b/plugins/codex-security/scripts/finalize_scan_contract.py
index 51259b0a4..9d6ec4f52 100644
--- a/plugins/codex-security/scripts/finalize_scan_contract.py
+++ b/plugins/codex-security/scripts/finalize_scan_contract.py
@@ -2699,7 +2699,7 @@ def write_sarif_projection(
def csv_cell(value: Any) -> Any:
if isinstance(value, str) and (
- value.startswith(("\t", "\r", "\n"))
+ value.startswith(("'", "\t", "\r", "\n"))
or value.lstrip().startswith(("=", "+", "-", "@", "=", "+", "-", "@"))
):
return f"'{value}"
diff --git a/sdk/typescript/README.md b/sdk/typescript/README.md
index 255e08ad6..4ef0de962 100644
--- a/sdk/typescript/README.md
+++ b/sdk/typescript/README.md
@@ -1889,6 +1889,11 @@ cancels the export. `export --help` lists the CLI options.
JSON preserves the sealed findings document. CSV marks findings as open,
omits local triage state, and cannot go to stdout when JSON output is requested.
+CSV escapes spreadsheet formula prefixes and literal leading apostrophes with
+an extra apostrophe; import removes that escape. Older CSV exports cannot
+distinguish some literal apostrophes from escapes. Use the JSON export when
+recovering those values from an older scan. Distinct CSV occurrence IDs are
+retained even when their finding IDs match, including when publishing CSV.
For CI, save output outside the checkout and set a severity threshold:
diff --git a/sdk/typescript/src/findings-import.ts b/sdk/typescript/src/findings-import.ts
index 4dcfb6fce..c5b639752 100644
--- a/sdk/typescript/src/findings-import.ts
+++ b/sdk/typescript/src/findings-import.ts
@@ -10,7 +10,9 @@ import type { Finding, FindingsDocument } from "./models.js";
export const CSV_TARGET_ID = "codex-security-csv-import";
export type FindingsImportFormat = "csv" | "json";
-const EXPORTED_CSV_ESCAPE = /^'(?:[\t\r\n]|\s*[=+\-@=+-@])/u;
+// Match the Python exporter's str.lstrip(), including its extra C0 separators.
+const EXPORTED_CSV_ESCAPE =
+ /^'(?:['\t\r\n]|[\p{White_Space}\u001c-\u001f]*[=+\-@=+-@])/u;
const csvFindingRowSchema = z
.object({
occurrence_id: z.string().regex(/^occ_[a-f0-9]{24}$/u, {
@@ -254,7 +256,6 @@ export function parseFindingsCsv(source: string): CsvFindingRow[] {
);
}
- const findingIds = new Set<string>();
const occurrenceIds = new Set<string>();
return rows.map((record, index) => {
const rowNumber = index + 2;
@@ -263,13 +264,9 @@ export function parseFindingsCsv(source: string): CsvFindingRow[] {
throw csvRowError(rowNumber, parsed.error.issues[0]!.message);
}
const row = parsed.data;
- if (findingIds.has(row.finding_id)) {
- throw csvRowError(rowNumber, "has a duplicate finding_id");
- }
if (occurrenceIds.has(row.occurrence_id)) {
throw csvRowError(rowNumber, "has a duplicate occurrence_id");
}
- findingIds.add(row.finding_id);
occurrenceIds.add(row.occurrence_id);
return row;
});
@@ -281,7 +278,7 @@ function decodeExportedCsvCell(value: string): string {
export function csvRowFinding(row: CsvFindingRow, scanId: string): Finding {
const ruleId = "import.csv";
- const anchor = row.finding_id;
+ const anchor = row.occurrence_id;
const fingerprint = `codex-security/v1:sha256:${sha256(
["codex-security/v1", CSV_TARGET_ID, ruleId, anchor, ""].join("\0"),
)}`;
diff --git a/sdk/typescript/tests-ts/cloud-publish.test.ts b/sdk/typescript/tests-ts/cloud-publish.test.ts
index 91ad0777b..fa034278f 100644
--- a/sdk/typescript/tests-ts/cloud-publish.test.ts
+++ b/sdk/typescript/tests-ts/cloud-publish.test.ts
@@ -214,6 +214,24 @@ describe("Cloud publication", () => {
expect(finding!.occurrenceId).toMatch(/^occ_[a-f0-9]{24}$/);
});
+ test("keeps separate CSV occurrences with the same source finding ID", async () => {
+ const second = csvRow.replace(
+ "occ_e79cb19591e696572a1c22be",
+ "occ_111111111111111111111111",
+ );
+ const path = await csvFixture(`${csvHeader}\n${csvRow}\n${second}\n`);
+ const result = await publishFindingsCsvToCloud(path, { dryRun: true });
+ expect(result.findingCount).toBe(2);
+ expect(
+ new Set(result.findings!.map((finding) => finding.occurrenceId)).size,
+ ).toBe(2);
+ expect(
+ new Set(result.findings!.map((finding) => finding.findingId)).size,
+ ).toBe(2);
+ const repeated = await publishFindingsCsvToCloud(path, { dryRun: true });
+ expect(repeated.findings).toEqual(result.findings);
+ });
+
test("accepts the candidate_id column from a Deep scan export", async () => {
const path = await csvFixture(
`${csvHeader.replace("finding_id,", "finding_id,candidate_id,")}\n${csvRow.replace("csf_852f90d6e1177502ff113d4a,", "csf_852f90d6e1177502ff113d4a,candidate-001,")}\n`,
diff --git a/sdk/typescript/tests-ts/findings-import.test.ts b/sdk/typescript/tests-ts/findings-import.test.ts
index 6c4ba695f..c62f8fc59 100644
--- a/sdk/typescript/tests-ts/findings-import.test.ts
+++ b/sdk/typescript/tests-ts/findings-import.test.ts
@@ -7,6 +7,7 @@ import {
} from "../src/findings-import.js";
import type { Finding, FindingsDocument } from "../src/models.js";
import { PLUGIN_ROOT } from "./plugin-root.js";
+import { PYTHON } from "./support/security-policy.js";
const BYTE_ORDER_MARK = "\uFEFF";
const CSV_SOURCE =
@@ -25,7 +26,121 @@ async function sourceDocument(): Promise<FindingsDocument> {
);
}
+function exportCsv(findings: Finding[]): string {
+ return execFileSync(
+ PYTHON,
+ [
+ "-c",
+ "import json, sys; sys.path.insert(0, sys.argv[1]); from finalize_scan_contract import build_csv_projection; sys.stdout.buffer.write(build_csv_projection(json.load(sys.stdin.buffer), {}))",
+ join(PLUGIN_ROOT, "scripts"),
+ ],
+ {
+ cwd: join(PLUGIN_ROOT, "scripts"),
+ input: JSON.stringify({ findings }),
+ encoding: "utf8",
+ },
+ );
+}
+
describe("findings import formats", () => {
+ test("retains distinct occurrences of a shared finding in CSV and JSON", async () => {
+ const {
+ findings: [finding],
+ } = await sourceDocument();
+ const rows = [
+ finding!,
+ { ...finding!, occurrenceId: "occ_111111111111111111111111" },
+ ];
+ for (const format of ["csv", "json"] as const) {
+ const source =
+ format === "csv" ? exportCsv(rows) : JSON.stringify({ findings: rows });
+ const parsed = await parseImportedFindings(source, format, PLUGIN_ROOT);
+ expect(
+ parsed.map(({ findingId, occurrenceId }) => [findingId, occurrenceId]),
+ ).toEqual(
+ rows.map(({ findingId, occurrenceId }) => [findingId, occurrenceId]),
+ );
+ expect(
+ new Set(
+ bindImportedFindings(parsed, format, "scan", "target").map(
+ (row) => row.occurrenceId,
+ ),
+ ).size,
+ ).toBe(2);
+ }
+ await expect(
+ parseImportedFindings(
+ exportCsv([finding!, finding!]),
+ "csv",
+ PLUGIN_ROOT,
+ ),
+ ).rejects.toThrow("duplicate occurrence_id");
+ });
+
+ test("round-trips literal apostrophes and spreadsheet escapes through the exporter", async () => {
+ const {
+ findings: [finding],
+ } = await sourceDocument();
+ const values = [
+ "'--no-verify' skips hooks",
+ "''=literal",
+ "'Literal",
+ "ordinary λ",
+ "=formula",
+ "\u0085=formula",
+ "\u001c=formula",
+ "\u001d=formula",
+ "\u001e=formula",
+ "\u001f=formula",
+ "\uFEFF=formula",
+ " +formula",
+ "-option",
+ "@owner",
+ "\tTabbed",
+ "\nNewline",
+ " =wide",
+ " +wide",
+ "-wide",
+ "@wide",
+ ];
+ for (const value of values) {
+ const source = {
+ ...finding!,
+ title: value,
+ summary: value,
+ remediation: value,
+ locations: [
+ {
+ path: value.replace(/[\u0000-\u001F]/gu, "") + ".ts",
+ startLine: 1,
+ },
+ ],
+ };
+ const [imported] = await parseImportedFindings(
+ exportCsv([source]),
+ "csv",
+ PLUGIN_ROOT,
+ );
+ expect(imported).toMatchObject({
+ title: source.title,
+ summary: source.summary,
+ remediation: source.remediation,
+ locations: source.locations,
+ });
+ }
+ });
+
+ test("retains a literal apostrophe before a non-whitespace byte order mark", async () => {
+ const title = "'\uFEFF=formula";
+ const [imported] = await parseImportedFindings(
+ CSV_SOURCE.replace("Reported CSV import issue", title),
+ "csv",
+ PLUGIN_ROOT,
+ );
+
+ expect(imported!.title).toBe(title);
+ });
+
test("accepts full findings documents and findings-service payloads", async () => {
const document = await sourceDocument();
document.findings[0]!.summary =
@@ -146,3 +261,4 @@ describe("findings import formats", () => {
).rejects.toThrow("duplicate occurrenceId");
});
});
+import { execFileSync } from "node:child_process";