fix: redact gh auth output in code queue

2026-05-23 04:22:42 +00:00
parent 5d5d38b9a1
commit 3b554e6154
7 changed files with 186 additions and 13 deletions
@@ -0,0 +1,121 @@
+import { mkdtempSync, readFileSync, rmSync } from "node:fs";
+import { tmpdir } from "node:os";
+import { join } from "node:path";
+import { appendOutput, configureTaskOutput, taskFullOutput } from "../src/components/microservices/code-queue/src/task-output";
+import { sanitizeTaskOutputText } from "../src/components/microservices/code-queue/src/output-redaction";
+import type { JsonValue, QueueTask } from "../src/components/microservices/code-queue/src/types";
+
+type JsonRecord = Record<string, unknown>;
+
+function assertCondition(condition: unknown, message: string, detail: unknown = {}): void {
+  if (!condition) throw new Error(`${message}: ${JSON.stringify(detail)}`);
+}
+
+function fixtureTask(): QueueTask {
+  const at = "2026-05-23T00:00:00.000Z";
+  return {
+    id: "codex_gh_auth_redaction_contract",
+    queueId: "default",
+    queueEnteredAt: at,
+    prompt: "redaction fixture",
+    basePrompt: "redaction fixture",
+    referenceTaskIds: [],
+    referenceInjection: null,
+    providerId: "D601",
+    cwd: "/workspace",
+    model: "gpt-5.5",
+    reasoningEffort: null,
+    executionMode: "default",
+    maxAttempts: 1,
+    status: "running",
+    createdAt: at,
+    updatedAt: at,
+    startedAt: at,
+    finishedAt: null,
+    readAt: null,
+    currentAttempt: 1,
+    currentMode: "initial",
+    codexThreadId: null,
+    activeTurnId: null,
+    finalResponse: "",
+    lastError: null,
+    lastJudge: null,
+    judgeFailCount: 0,
+    promptHistory: [],
+    output: [],
+    events: [],
+    attempts: [],
+    cancelRequested: false,
+    nextPrompt: null,
+    nextMode: null,
+  };
+}
+
+function assertNoTokenFragments(value: string, label: string): void {
+  assertCondition(!/\bgh[pousr]_[A-Za-z0-9_]{6,}\b/u.test(value), `${label} must redact gh token-like values`, value);
+  assertCondition(!/\bgithub_pat_[A-Za-z0-9_]{6,}\b/u.test(value), `${label} must redact GitHub PAT-like values`, value);
+  assertCondition(!/Token:\s*\S+/iu.test(value), `${label} must not expose raw gh auth token lines`, value);
+  assertCondition(!/Token scopes?:\s*\S+/iu.test(value), `${label} must not expose raw gh auth scope lines`, value);
+}
+
+export function runCodeQueueGhAuthRedactionContract(): JsonRecord {
+  const tmp = mkdtempSync(join(tmpdir(), "code-queue-gh-auth-redaction-"));
+  const task = fixtureTask();
+  let seq = 0;
+  try {
+    configureTaskOutput({
+      config: { maxInMemoryOutputRecords: 1000, outputArchiveDir: tmp },
+      allocateSeq: () => {
+        seq += 1;
+        return seq;
+      },
+      errorToJson: (error: unknown): JsonValue => error instanceof Error ? { message: error.message } : String(error),
+      logger: () => undefined,
+      markTaskDirty: () => undefined,
+      nowIso: () => "2026-05-23T00:00:01.000Z",
+      schedulePersistState: () => undefined,
+    });
+
+    const rawGhAuth = [
+      "github.com",
+      "  \u2713 Logged in to github.com account example (keyring)",
+      "  - Active account: true",
+      "  - Git operations protocol: ssh",
+      "  - Token: ghp_abcdef1234567890abcdef1234567890",
+      "  - Token scopes: 'repo', 'read:org'",
+      "generic token=github_pat_abcdef1234567890abcdef1234567890",
+    ].join("\n");
+    const sanitized = sanitizeTaskOutputText(rawGhAuth);
+    assertNoTokenFragments(sanitized, "direct sanitizer output");
+    assertCondition(sanitized.includes("[redacted gh auth status line]"), "sanitizer must redact gh auth status lines", sanitized);
+    assertCondition(sanitized.includes("bun scripts/cli.ts gh auth status"), "sanitizer must include UniDesk gh wrapper hint", sanitized);
+
+    appendOutput(task, "command", `${rawGhAuth}\n`, "item/commandExecution/outputDelta", "call-gh-auth", true);
+    const retained = JSON.stringify(task.output);
+    const archived = readFileSync(join(tmp, `${task.id}.jsonl`), "utf8");
+    const full = JSON.stringify(taskFullOutput(task));
+    assertNoTokenFragments(retained, "retained output");
+    assertNoTokenFragments(archived, "archived output");
+    assertNoTokenFragments(full, "full output replay");
+    assertCondition(full.includes("bun scripts/cli.ts gh auth status"), "full output replay must keep wrapper hint", full);
+
+    return {
+      ok: true,
+      checks: [
+        "raw gh auth status token and scope lines are redacted",
+        "token-like GitHub values are redacted before retained output persistence",
+        "output archive replay remains redacted",
+        "redacted output nudges runner toward bun scripts/cli.ts gh auth status",
+      ],
+    };
+  } finally {
+    rmSync(tmp, { recursive: true, force: true });
+  }
+}
+
+try {
+  process.stdout.write(`${JSON.stringify(runCodeQueueGhAuthRedactionContract(), null, 2)}\n`);
+} catch (error) {
+  process.stderr.write(`${error instanceof Error ? error.stack ?? error.message : String(error)}\n`);
+  process.exit(1);
+}