@openai/codex-security 0.1.10 → 0.1.12

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (99) hide show
  1. package/README.md +72 -41
  2. package/_bundled_plugin/.codex-plugin/plugin.json +1 -1
  3. package/_bundled_plugin/.mcp.json +5 -1
  4. package/_bundled_plugin/examples/completed-scan/coverage.json +1 -1
  5. package/_bundled_plugin/examples/completed-scan/scan-manifest.json +2 -2
  6. package/_bundled_plugin/mcp/server.mjs.br.part-000 +0 -0
  7. package/_bundled_plugin/mcp/server.mjs.br.part-001 +0 -0
  8. package/_bundled_plugin/preflight/capability-profiles.toml +1 -34
  9. package/_bundled_plugin/references/config-preflight.md +11 -13
  10. package/_bundled_plugin/references/core-scan.md +94 -0
  11. package/_bundled_plugin/references/final-report.md +6 -6
  12. package/_bundled_plugin/references/scan-artifacts.md +15 -14
  13. package/_bundled_plugin/references/scan-prologue.md +27 -0
  14. package/_bundled_plugin/schemas/tools/deep-reducer.schema.json +1 -106
  15. package/_bundled_plugin/scripts/config_preflight.py +40 -3
  16. package/_bundled_plugin/scripts/deep_scan_config.py +29 -6
  17. package/_bundled_plugin/scripts/deep_scan_workbench.py +132 -34
  18. package/_bundled_plugin/scripts/finalize_scan_contract.py +44 -224
  19. package/_bundled_plugin/scripts/finding_preview.py +122 -6
  20. package/_bundled_plugin/scripts/generate_in_scope_files.py +77 -31
  21. package/_bundled_plugin/scripts/generate_rank_input.py +58 -7
  22. package/_bundled_plugin/scripts/normalize_candidates.py +5 -5
  23. package/_bundled_plugin/scripts/rank_preview.py +4 -0
  24. package/_bundled_plugin/scripts/report_projection.py +40 -20
  25. package/_bundled_plugin/scripts/workbench_cli.py +5 -0
  26. package/_bundled_plugin/scripts/workbench_db.py +304 -32
  27. package/_bundled_plugin/scripts/workbench_progress.py +0 -5
  28. package/_bundled_plugin/scripts/workbench_scan_usage.py +43 -8
  29. package/_bundled_plugin/scripts/workbench_schema.py +16 -20
  30. package/_bundled_plugin/scripts/workbench_source_excerpt.py +1 -5
  31. package/_bundled_plugin/scripts/workbench_target.py +82 -0
  32. package/_bundled_plugin/skills/attack-path-analysis/SKILL.md +8 -11
  33. package/_bundled_plugin/skills/deep-security-scan/SKILL.md +24 -108
  34. package/_bundled_plugin/skills/security-scan/SKILL.md +5 -88
  35. package/_bundled_plugin/skills/threat-model/SKILL.md +4 -2
  36. package/_bundled_plugin/skills/track-findings/SKILL.md +1 -1
  37. package/_bundled_plugin/skills/triage-finding/SKILL.md +5 -6
  38. package/_bundled_plugin/skills/triage-finding/references/github-rest-intake.md +6 -7
  39. package/_bundled_plugin/skills/validation/SKILL.md +9 -12
  40. package/_bundled_plugin/skills/vulnerability-writeup/SKILL.md +3 -2
  41. package/_bundled_plugin/skills/vulnerability-writeup/references/report-format.md +1 -1
  42. package/dist/api.d.ts +2 -0
  43. package/dist/api.d.ts.map +1 -1
  44. package/dist/api.js +393 -145
  45. package/dist/api.js.map +1 -1
  46. package/dist/auth.d.ts +2 -3
  47. package/dist/auth.d.ts.map +1 -1
  48. package/dist/auth.js +59 -270
  49. package/dist/auth.js.map +1 -1
  50. package/dist/bulk-scan-discovery.js +14 -2
  51. package/dist/bulk-scan-discovery.js.map +1 -1
  52. package/dist/cli.d.ts +6 -3
  53. package/dist/cli.d.ts.map +1 -1
  54. package/dist/cli.js +202 -46
  55. package/dist/cli.js.map +1 -1
  56. package/dist/config.d.ts.map +1 -1
  57. package/dist/config.js +1 -0
  58. package/dist/config.js.map +1 -1
  59. package/dist/contract.js +7 -5
  60. package/dist/contract.js.map +1 -1
  61. package/dist/cost.d.ts +1 -0
  62. package/dist/cost.d.ts.map +1 -1
  63. package/dist/cost.js +33 -2
  64. package/dist/cost.js.map +1 -1
  65. package/dist/knowledge-base.d.ts.map +1 -1
  66. package/dist/knowledge-base.js +12 -4
  67. package/dist/knowledge-base.js.map +1 -1
  68. package/dist/multiscan.d.ts +1 -0
  69. package/dist/multiscan.d.ts.map +1 -1
  70. package/dist/multiscan.js +77 -17
  71. package/dist/multiscan.js.map +1 -1
  72. package/dist/runtime.d.ts +9 -6
  73. package/dist/runtime.d.ts.map +1 -1
  74. package/dist/runtime.js +140 -326
  75. package/dist/runtime.js.map +1 -1
  76. package/dist/scan-comparison.js +1 -1
  77. package/dist/scan-comparison.js.map +1 -1
  78. package/dist/scan-dashboard.d.ts +5 -1
  79. package/dist/scan-dashboard.d.ts.map +1 -1
  80. package/dist/scan-dashboard.js +65 -16
  81. package/dist/scan-dashboard.js.map +1 -1
  82. package/dist/scan-logs.d.ts +2 -0
  83. package/dist/scan-logs.d.ts.map +1 -1
  84. package/dist/scan-logs.js +97 -37
  85. package/dist/scan-logs.js.map +1 -1
  86. package/dist/targets.d.ts +1 -0
  87. package/dist/targets.d.ts.map +1 -1
  88. package/dist/targets.js +35 -4
  89. package/dist/targets.js.map +1 -1
  90. package/dist/trusted-executable.js +1 -1
  91. package/dist/trusted-executable.js.map +1 -1
  92. package/dist/version.d.ts +3 -3
  93. package/dist/version.d.ts.map +1 -1
  94. package/dist/version.js +6 -4
  95. package/dist/version.js.map +1 -1
  96. package/package.json +4 -4
  97. package/_bundled_plugin/references/shared-hard-rules.md +0 -15
  98. package/_bundled_plugin/scripts/validate_report_format.py +0 -138
  99. package/_bundled_plugin/skills/security-scan/references/repository-wide-scan.md +0 -33
@@ -97,7 +97,7 @@ from workbench_schema import (
97
97
  from workbench_schema import (
98
98
  sql_statements as sql_statements,
99
99
  )
100
- from workbench_source_excerpt import finding_source_excerpt
100
+ from workbench_source_excerpt import finding_source_excerpt, safe_source_path
101
101
  from workbench_target import (
102
102
  clean_worktree_content_digest,
103
103
  copy_directory_excluding,
@@ -1136,6 +1136,289 @@ def complete_scan(
1136
1136
  )
1137
1137
 
1138
1138
 
1139
+ def complete_budget_exhausted_scan(
1140
+ connection: sqlite3.Connection, args: argparse.Namespace
1141
+ ) -> dict[str, Any]:
1142
+ scan_id = require_uuid(args.scan_id, "scan-id")
1143
+ cost_json = parse_scan_cost(args.cost_json)
1144
+ if cost_json is None:
1145
+ raise SystemExit("Budget-exhausted scan completion requires the measured scan cost.")
1146
+ with scan_completion_lock(scan_id):
1147
+ scan = require_scan(connection, scan_id)
1148
+ if scan["status"] != "running" or scan["mode"] != "deep" or scan["recipe_json"] is None:
1149
+ raise SystemExit("Only a running CLI Deep Scan can complete after its cost limit.")
1150
+ recipe = json.loads(scan["recipe_json"], parse_constant=reject_non_finite_json)
1151
+ if not isinstance(recipe, dict) or recipe.get("mode") != "deep":
1152
+ raise SystemExit("Budget-exhausted scan completion requires a Deep Scan launch recipe.")
1153
+ cost = json.loads(cost_json)
1154
+ measured = cost.get("cost", cost)
1155
+ limit = recipe.get("maxCostUsd")
1156
+ if (
1157
+ not isinstance(limit, (int, float))
1158
+ or isinstance(limit, bool)
1159
+ or not isinstance(measured, dict)
1160
+ or measured.get("estimatedUsd", 0) <= limit
1161
+ ):
1162
+ raise SystemExit("Deep Scan has not exceeded its configured cost limit.")
1163
+ run = connection.execute(
1164
+ "SELECT status, terminal_reason, manifest_path FROM deep_scan_runs WHERE scan_id = ?",
1165
+ (scan_id,),
1166
+ ).fetchone()
1167
+ if (
1168
+ run is None
1169
+ or run["status"] != "succeeded"
1170
+ or run["terminal_reason"] not in {"saturated", "capped"}
1171
+ or not run["manifest_path"]
1172
+ ):
1173
+ raise SystemExit(
1174
+ "Budget-exhausted scan completion requires successfully completed Deep Scan "
1175
+ "discovery."
1176
+ )
1177
+ scan_dir = require_canonical_scan_directory(Path(scan["scan_dir"]))
1178
+ candidates = (
1179
+ []
1180
+ if run["manifest_path"] == str(scan_dir / "scan-manifest.json")
1181
+ else budget_exhausted_candidates(scan, scan_dir)
1182
+ )
1183
+ warning = optional_text(args.message, maximum=2400)
1184
+ if warning is None:
1185
+ warning = (
1186
+ f"Deep Scan reached its cost limit after an estimated "
1187
+ f"${measured['estimatedUsd']:.6g}; completed discovery was preserved."
1188
+ )
1189
+ budget_exhausted_draft(scan, scan_dir, candidates, warning)
1190
+ warnings = json.loads(scan["completion_warnings_json"])
1191
+ if warning not in warnings:
1192
+ connection.execute(
1193
+ "UPDATE scans SET completion_warnings_json = ? WHERE id = ? AND status = 'running'",
1194
+ (json.dumps([*warnings, warning]), scan_id),
1195
+ )
1196
+ connection.commit()
1197
+ return complete_scan_locked(connection, scan_id, None, cost_json)
1198
+
1199
+
1200
+ def budget_exhausted_candidates(scan: sqlite3.Row, scan_dir: Path) -> list[dict[str, Any]]:
1201
+ artifacts = deep_scan.canonical_discovery_artifacts(scan)
1202
+ ledger = Path(artifacts["candidateLedgerPath"])
1203
+ try:
1204
+ inventory = Path(artifacts["inScopeFilesPath"])
1205
+ inventory_descriptor = open_scan_local_file_descriptor(
1206
+ scan_dir,
1207
+ inventory.relative_to(scan_dir).as_posix(),
1208
+ "Canonical Deep Scan in-scope inventory",
1209
+ )
1210
+ with os.fdopen(inventory_descriptor, "rb") as source:
1211
+ lines = re.split(r"\r?\n", source.read().decode("utf-8"))
1212
+ in_scope = {re.sub(r"^(?:\./)+", "", line) for line in lines if line}
1213
+ descriptor = open_scan_local_file_descriptor(
1214
+ scan_dir,
1215
+ ledger.relative_to(scan_dir).as_posix(),
1216
+ "Canonical Deep Scan candidate ledger",
1217
+ )
1218
+ with os.fdopen(descriptor, "r", encoding="utf-8") as source:
1219
+ candidates = [
1220
+ json.loads(line, parse_constant=reject_non_finite_json)
1221
+ for line in source
1222
+ if line.strip()
1223
+ ]
1224
+ except (ContractError, OSError, UnicodeError, ValueError) as exc:
1225
+ raise SystemExit(f"Canonical Deep Scan candidate ledger is invalid: {exc}") from exc
1226
+
1227
+ candidate_ids: set[str] = set()
1228
+ for candidate in candidates:
1229
+ if not isinstance(candidate, dict):
1230
+ raise SystemExit("Canonical Deep Scan candidate ledger rows must be objects.")
1231
+ candidate_id = candidate.get("candidate_id")
1232
+ locations = candidate.get("locations")
1233
+ if (
1234
+ not isinstance(candidate_id, str)
1235
+ or not candidate_id.strip()
1236
+ or candidate_id in {".", ".."}
1237
+ or "/" in candidate_id
1238
+ or "\\" in candidate_id
1239
+ or candidate_id in candidate_ids
1240
+ or not isinstance(candidate.get("summary"), str)
1241
+ or not candidate["summary"].strip()
1242
+ or not isinstance(candidate.get("evidence"), str)
1243
+ or not candidate["evidence"].strip()
1244
+ or not isinstance(locations, list)
1245
+ or not locations
1246
+ ):
1247
+ raise SystemExit("Canonical Deep Scan candidate ledger contains an invalid candidate.")
1248
+ candidate_ids.add(candidate_id)
1249
+ for location in locations:
1250
+ if not isinstance(location, dict):
1251
+ raise SystemExit("Canonical Deep Scan candidate location must be an object.")
1252
+ path = location.get("path")
1253
+ if (
1254
+ not isinstance(path, str)
1255
+ or not path
1256
+ or "\\" in path
1257
+ or "\x00" in path
1258
+ or re.match(r"^[A-Za-z]:", path)
1259
+ or PurePosixPath(path).is_absolute()
1260
+ or any(part in {".", "..", ""} for part in path.split("/"))
1261
+ ):
1262
+ raise SystemExit(
1263
+ "Canonical Deep Scan candidate location must be repository-relative."
1264
+ )
1265
+ if not any(location["path"] in in_scope for location in locations):
1266
+ raise SystemExit(
1267
+ "Canonical Deep Scan candidate must include a location in its in-scope inventory."
1268
+ )
1269
+ return candidates
1270
+
1271
+
1272
+ def budget_exhausted_draft(
1273
+ scan: sqlite3.Row,
1274
+ scan_dir: Path,
1275
+ candidates: list[dict[str, Any]],
1276
+ warning: str,
1277
+ ) -> None:
1278
+ documents: dict[str, dict[str, Any]] = {}
1279
+ for name in ("scan-manifest.json", "findings.json", "coverage.json"):
1280
+ path = artifact_path(scan_dir, name, required=False)
1281
+ if path is not None:
1282
+ documents[name] = read_json_object(path)
1283
+ if documents and len(documents) != 3:
1284
+ raise SystemExit("Budget-exhausted scan contains an incomplete canonical scan draft.")
1285
+
1286
+ if documents:
1287
+ manifest = documents["scan-manifest.json"]
1288
+ findings = documents["findings.json"]
1289
+ coverage = documents["coverage.json"]
1290
+ if not isinstance(manifest.get("scan"), dict) or not isinstance(
1291
+ findings.get("findings"), list
1292
+ ):
1293
+ raise SystemExit("Budget-exhausted scan contains an invalid canonical scan draft.")
1294
+ for key in ("surfaces", "explicitExclusions", "deferred"):
1295
+ if not isinstance(coverage.get(key), list):
1296
+ raise SystemExit("Budget-exhausted scan contains invalid canonical coverage.")
1297
+ if manifest["scan"].get("sealedAt") is not None or manifest["scan"].get("artifacts"):
1298
+ raise SystemExit("Budget-exhausted scan cannot replace an already sealed scan draft.")
1299
+ else:
1300
+ contract = scan_contract(scan)
1301
+ target_contract = contract["target"]
1302
+ target: dict[str, Any] = {
1303
+ "kind": target_contract["allowedKinds"][0],
1304
+ "targetId": target_contract["targetId"],
1305
+ "displayName": target_contract["displayName"],
1306
+ }
1307
+ if scan["target_revision"] != "unversioned":
1308
+ target["revision"] = scan["target_revision"]
1309
+ if "requiredSnapshotDigest" in target_contract:
1310
+ target["snapshotDigest"] = target_contract["requiredSnapshotDigest"]
1311
+ manifest = {
1312
+ "scan": {
1313
+ "target": target,
1314
+ "scope": {"limitations": [warning], "validationMode": "incomplete"},
1315
+ }
1316
+ }
1317
+ findings = {"findings": []}
1318
+ coverage = {
1319
+ "completeness": "partial",
1320
+ "inventoryStrategy": (
1321
+ "scoped_path" if expected_coverage_mode(scan) == "scoped_path" else "repository"
1322
+ ),
1323
+ "surfaces": [],
1324
+ "explicitExclusions": [],
1325
+ "deferred": [],
1326
+ }
1327
+
1328
+ findings_by_candidate = {
1329
+ candidate_id
1330
+ for finding in findings["findings"]
1331
+ if isinstance(finding, dict)
1332
+ and isinstance(candidate_id := finding_candidate_id(finding), str)
1333
+ }
1334
+ existing_deferred = {
1335
+ item.get("candidateId", item.get("id"))
1336
+ for item in coverage["deferred"]
1337
+ if isinstance(item, dict)
1338
+ and isinstance(item.get("candidateId", item.get("id")), str)
1339
+ }
1340
+ existing_surfaces = {
1341
+ item.get("id")
1342
+ for item in coverage["surfaces"]
1343
+ if isinstance(item, dict) and isinstance(item.get("id"), str)
1344
+ }
1345
+ for candidate in candidates:
1346
+ candidate_id = candidate["candidate_id"]
1347
+ if candidate_id in findings_by_candidate or candidate_id in existing_deferred:
1348
+ continue
1349
+ paths = list(dict.fromkeys(location["path"] for location in candidate["locations"]))
1350
+ surface_id = f"candidate-{candidate_id}"
1351
+ validation = candidate.get("validation")
1352
+ validation = validation.get("disposition") if isinstance(validation, dict) else None
1353
+ attack = candidate.get("attack_path")
1354
+ attack = attack.get("decision") if isinstance(attack, dict) else None
1355
+ disposition = (
1356
+ "needs_follow_up"
1357
+ if validation == "deferred" or attack == "deferred"
1358
+ else "not_applicable"
1359
+ if validation == "not_applicable"
1360
+ else "rejected"
1361
+ if validation == "suppressed" or attack == "ignore"
1362
+ else "needs_follow_up"
1363
+ )
1364
+ if surface_id not in existing_surfaces:
1365
+ coverage["surfaces"].append(
1366
+ {
1367
+ "id": surface_id,
1368
+ "label": candidate["summary"],
1369
+ "disposition": disposition,
1370
+ "notes": candidate["evidence"],
1371
+ "receiptRefs": [],
1372
+ }
1373
+ )
1374
+ existing_surfaces.add(surface_id)
1375
+ if disposition != "needs_follow_up":
1376
+ continue
1377
+ coverage["deferred"].append(
1378
+ {
1379
+ "id": candidate_id,
1380
+ "candidateId": candidate_id,
1381
+ "reason": (
1382
+ "Validation was deferred because the scan reached its cost limit: "
1383
+ f"{candidate['summary']}. Evidence: {candidate['evidence']}"
1384
+ ),
1385
+ "paths": paths,
1386
+ "surfaceIds": [surface_id],
1387
+ }
1388
+ )
1389
+ if not any(
1390
+ isinstance(item, dict)
1391
+ and isinstance(reason := item.get("reason"), str)
1392
+ and (
1393
+ reason == "Validation was deferred because the scan reached its cost limit."
1394
+ or reason.startswith(
1395
+ "Validation was deferred because the scan reached its cost limit: "
1396
+ )
1397
+ )
1398
+ for item in coverage["deferred"]
1399
+ ):
1400
+ coverage["deferred"].append(
1401
+ {
1402
+ "id": "scan-cost-limit",
1403
+ "reason": "Validation was deferred because the scan reached its cost limit.",
1404
+ }
1405
+ )
1406
+ coverage["completeness"] = "partial"
1407
+ for name, payload in (
1408
+ ("findings.json", findings),
1409
+ ("coverage.json", coverage),
1410
+ ("scan-manifest.json", manifest),
1411
+ ):
1412
+ try:
1413
+ write_scan_local_bytes(
1414
+ scan_dir,
1415
+ name,
1416
+ (json.dumps(payload, allow_nan=False, indent=2, sort_keys=True) + "\n").encode(),
1417
+ )
1418
+ except (ContractError, OSError, TypeError, ValueError) as exc:
1419
+ raise SystemExit(f"Budget-exhausted scan draft could not be saved: {exc}") from exc
1420
+
1421
+
1139
1422
  def complete_scan_locked(
1140
1423
  connection: sqlite3.Connection,
1141
1424
  scan_id: str,
@@ -1168,15 +1451,17 @@ def complete_scan_locked(
1168
1451
  claim_token,
1169
1452
  error_message="Scan completion is owned by another continuation.",
1170
1453
  )
1171
- if scan["recipe_json"] is None:
1172
- deep_scan.require_deep_scan_ready_for_parent_completion(connection, scan)
1454
+ deep_scan.require_deep_scan_ready_for_parent_completion(connection, scan)
1173
1455
  warnings = json.loads(scan["completion_warnings_json"])
1174
1456
  target_warnings: list[str] = []
1175
- warning = scan_target_warning(scan)
1176
- if warning is not None:
1177
- target_warnings.append(warning)
1178
- if warning not in warnings:
1179
- warnings.append(warning)
1457
+
1458
+ def add_warning() -> None:
1459
+ if (warning := scan_target_warning(scan)) is not None:
1460
+ for items in (target_warnings, warnings):
1461
+ if warning not in items:
1462
+ items.append(warning)
1463
+
1464
+ add_warning()
1180
1465
  scan_dir = require_canonical_scan_directory(Path(scan["scan_dir"]))
1181
1466
  completion_timestamp = now()
1182
1467
  completion_binding = workbench_completion_binding(scan, completion_timestamp)
@@ -1214,12 +1499,7 @@ def complete_scan_locked(
1214
1499
  completion_binding=completion_binding,
1215
1500
  completion_warnings=warnings,
1216
1501
  )
1217
- warning = scan_target_warning(scan)
1218
- if warning is not None:
1219
- if warning not in target_warnings:
1220
- target_warnings.append(warning)
1221
- if warning not in warnings:
1222
- warnings.append(warning)
1502
+ add_warning()
1223
1503
  manifest, findings, _ = _write_prepared_scan_finalization(prepared)
1224
1504
  except ContractError as exc:
1225
1505
  raise SystemExit(str(exc)) from exc
@@ -1262,8 +1542,7 @@ def complete_scan_locked(
1262
1542
  return scan_context(connection, scan["id"])
1263
1543
  if scan["status"] != "running":
1264
1544
  raise SystemExit("Only a running scan can be completed.")
1265
- if scan["recipe_json"] is None:
1266
- deep_scan.require_deep_scan_ready_for_parent_completion(connection, scan)
1545
+ deep_scan.require_deep_scan_ready_for_parent_completion(connection, scan)
1267
1546
  handoff.require_current_continuation(
1268
1547
  scan,
1269
1548
  claim_token,
@@ -2628,10 +2907,15 @@ def scan_context(
2628
2907
  occurrence_id: str | None = None,
2629
2908
  ) -> dict[str, Any]:
2630
2909
  scan = require_scan(connection, scan_id)
2910
+ workspace = workspace_state(connection, scan["workspace_id"], result_scan_id=scan["id"])
2631
2911
  context = {
2632
2912
  "otherRunningDeepScans": deep_scan.other_running_deep_scans(connection, scan["id"]),
2633
- "scan": scan_result(connection, scan, occurrence_id=occurrence_id),
2634
- "workspace": workspace_state(connection, scan["workspace_id"], result_scan_id=scan["id"]),
2913
+ "scan": (
2914
+ workspace["results"]
2915
+ if occurrence_id is None
2916
+ else scan_result(connection, scan, occurrence_id=occurrence_id)
2917
+ ),
2918
+ "workspace": workspace,
2635
2919
  }
2636
2920
  if scan["recipe_json"] is not None:
2637
2921
  context["parentScanId"] = scan["parent_scan_id"]
@@ -3185,20 +3469,6 @@ def patch_artifact_preview(
3185
3469
  }
3186
3470
 
3187
3471
 
3188
- def safe_source_path(target: Path, relative_path: str) -> Path | None:
3189
- if "\\" in relative_path:
3190
- return None
3191
- parsed = PurePosixPath(relative_path)
3192
- if parsed.is_absolute() or ".." in parsed.parts:
3193
- return None
3194
- try:
3195
- path = (target / parsed.as_posix()).resolve()
3196
- path.relative_to(target)
3197
- except (OSError, RuntimeError, ValueError):
3198
- return None
3199
- return path
3200
-
3201
-
3202
3472
  def available_artifact_path(scan_dir: Path, candidate: Path) -> Path | None:
3203
3473
  try:
3204
3474
  resolved_scan_dir = require_canonical_scan_directory(scan_dir)
@@ -3400,6 +3670,8 @@ def main() -> None:
3400
3670
  result = complete_scan(
3401
3671
  connection, args, prepare_only=args.command == "prepare-scan-completion"
3402
3672
  )
3673
+ elif args.command == "complete-budget-exhausted-scan":
3674
+ result = complete_budget_exhausted_scan(connection, args)
3403
3675
  elif args.command == "cancel-scan":
3404
3676
  result = cancel_scan(connection, args)
3405
3677
  elif args.command == "fail-scan":
@@ -13,7 +13,6 @@ from workbench.handoff import require_current_continuation
13
13
  from workbench_constants import PHASES
14
14
  from workbench_validation import optional_text, require_uuid, user_text
15
15
 
16
- MAX_PREFLIGHT_ISSUES_JSON_BYTES = 64 * 1024
17
16
  MAX_PREFLIGHT_ISSUES = 32
18
17
 
19
18
 
@@ -33,10 +32,6 @@ def _preflight_issue_text(value: Any, maximum: int, label: str) -> str:
33
32
  def preflight_issues_json(value: str | None) -> str | None:
34
33
  if value is None:
35
34
  return None
36
- if len(value.encode("utf-8")) > MAX_PREFLIGHT_ISSUES_JSON_BYTES:
37
- raise SystemExit(
38
- f"Preflight issues must be no larger than {MAX_PREFLIGHT_ISSUES_JSON_BYTES} bytes."
39
- )
40
35
  try:
41
36
  payload = json.loads(value)
42
37
  except json.JSONDecodeError as exc:
@@ -189,10 +189,18 @@ def _codex_state_database() -> Path | None:
189
189
  return path.resolve() if path.is_file() and os.access(path, os.R_OK) else None
190
190
 
191
191
  configured_home = os.environ.get("CODEX_HOME", "").strip()
192
- codex_home = Path(configured_home).expanduser() if configured_home else Path.home() / ".codex"
192
+ codex_home = (
193
+ Path(configured_home).expanduser()
194
+ if configured_home
195
+ else Path.home() / ".codex"
196
+ )
193
197
  configured_sqlite_home = os.environ.get("CODEX_SQLITE_HOME", "").strip()
194
198
  search_roots = [
195
- *([Path(configured_sqlite_home).expanduser()] if configured_sqlite_home else []),
199
+ *(
200
+ [Path(configured_sqlite_home).expanduser()]
201
+ if configured_sqlite_home
202
+ else []
203
+ ),
196
204
  codex_home,
197
205
  codex_home / "sqlite",
198
206
  ]
@@ -329,7 +337,9 @@ def _require_state_columns(
329
337
  }
330
338
  columns = {str(row["name"]) for row in connection.execute(statements[table])}
331
339
  if not required.issubset(columns):
332
- raise ValueError("Codex state graph does not expose the required thread columns.")
340
+ raise ValueError(
341
+ "Codex state graph does not expose the required thread columns."
342
+ )
333
343
 
334
344
 
335
345
  def _rollout_path(value: object) -> Path | None:
@@ -375,7 +385,9 @@ def _read_rollout_usage(
375
385
  event = json.loads(raw_line)
376
386
  except (UnicodeError, ValueError):
377
387
  if line_number == 1:
378
- raise ValueError("The rollout session metadata is unreadable.") from None
388
+ raise ValueError(
389
+ "The rollout session metadata is unreadable."
390
+ ) from None
379
391
  if boundary_reached:
380
392
  warnings.add("rollout_record_invalid")
381
393
  continue
@@ -418,7 +430,10 @@ def _read_rollout_usage(
418
430
  warnings.add("thread_outside_scan_window")
419
431
  return total, warnings
420
432
  boundary_reached = True
421
- elif event.get("type") == "event_msg" and payload.get("type") == "token_count":
433
+ elif (
434
+ event.get("type") == "event_msg"
435
+ and payload.get("type") == "token_count"
436
+ ):
422
437
  inherited_usage = _token_snapshot(payload)
423
438
  if inherited_usage is not None:
424
439
  previous = inherited_usage
@@ -497,10 +512,25 @@ def _token_snapshot(payload: Mapping[str, Any]) -> dict[str, int] | None:
497
512
  usage = info.get("total_token_usage")
498
513
  if not isinstance(usage, dict):
499
514
  return None
515
+ cache_write = usage.get(
516
+ "cache_write_input_tokens", usage.get("cache_write_tokens", 0)
517
+ )
518
+ legacy_cache_write = usage.get("cache_write_tokens")
519
+ input_tokens = usage.get("input_tokens")
520
+ cached_input_tokens = usage.get("cached_input_tokens", 0)
521
+ if (
522
+ cache_write == 0
523
+ and type(legacy_cache_write) is int
524
+ and legacy_cache_write > 0
525
+ and type(input_tokens) is int
526
+ and type(cached_input_tokens) is int
527
+ and cached_input_tokens + legacy_cache_write <= input_tokens
528
+ ):
529
+ cache_write = legacy_cache_write
500
530
  result: dict[str, int] = {}
501
531
  for source_key, result_key in TOKEN_FIELDS.items():
502
532
  value = (
503
- usage.get(source_key, usage.get("cache_write_tokens", 0))
533
+ cache_write
504
534
  if source_key == "cache_write_input_tokens"
505
535
  else usage.get(source_key, 0)
506
536
  )
@@ -511,7 +541,10 @@ def _token_snapshot(payload: Mapping[str, Any]) -> dict[str, int] | None:
511
541
  ):
512
542
  return None
513
543
  result[result_key] = value
514
- if result["cachedInputTokens"] + result["cacheWriteInputTokens"] > result["inputTokens"]:
544
+ if (
545
+ result["cachedInputTokens"] + result["cacheWriteInputTokens"]
546
+ > result["inputTokens"]
547
+ ):
515
548
  return None
516
549
  result["totalTokens"] = result["inputTokens"] + result["outputTokens"]
517
550
  return result
@@ -540,7 +573,9 @@ def _timestamp(value: object) -> datetime | None:
540
573
  return parsed.astimezone(timezone.utc) if parsed.tzinfo is not None else None
541
574
 
542
575
 
543
- def _unavailable_usage(reason: str, *, warnings: set[str] | None = None) -> dict[str, Any]:
576
+ def _unavailable_usage(
577
+ reason: str, *, warnings: set[str] | None = None
578
+ ) -> dict[str, Any]:
544
579
  return {
545
580
  "coverage": "unavailable",
546
581
  "source": "codex_rollout",
@@ -635,6 +635,14 @@ MIGRATIONS = (
635
635
  SET stop_after_consecutive_errors = stop_after_no_new;
636
636
  """,
637
637
  ),
638
+ (
639
+ 28,
640
+ "persist deep scan discovery time limit",
641
+ """
642
+ ALTER TABLE deep_scan_runs
643
+ ADD COLUMN max_time_hours REAL NOT NULL DEFAULT 96;
644
+ """,
645
+ ),
638
646
  )
639
647
 
640
648
 
@@ -689,6 +697,13 @@ def apply_migrations(
689
697
  "completion_warnings_json",
690
698
  "TEXT NOT NULL DEFAULT '[]'",
691
699
  )
700
+ elif version == 28:
701
+ add_column_if_missing(
702
+ connection,
703
+ "deep_scan_runs",
704
+ "max_time_hours",
705
+ "REAL NOT NULL DEFAULT 96",
706
+ )
692
707
  continue
693
708
  if version == 6:
694
709
  repair_thread_scoped_workspaces_migration(connection)
@@ -729,7 +744,6 @@ def normalize_pre_release_execution_profile_migrations(
729
744
  12: "scan continuation threads",
730
745
  }
731
746
  model_migration_name = "persist scan model settings"
732
- warnings_migration_name = "persist scan completion warnings"
733
747
  if execution_migrations.get(25) == "dynamic scan execution profiles":
734
748
  connection.execute(
735
749
  "UPDATE schema_migrations SET name = ? WHERE version = 25 AND name = ?",
@@ -737,9 +751,6 @@ def normalize_pre_release_execution_profile_migrations(
737
751
  )
738
752
  execution_migrations[25] = model_migration_name
739
753
  has_legacy_profile_history = execution_migrations.get(11) == legacy_names[11]
740
- has_public_warnings_history = (
741
- execution_migrations.get(25) == warnings_migration_name
742
- )
743
754
 
744
755
  scan_columns = {
745
756
  row["name"] for row in connection.execute("PRAGMA table_info(scans)")
@@ -750,11 +761,7 @@ def normalize_pre_release_execution_profile_migrations(
750
761
  has_legacy_profile_columns = (
751
762
  "execution_model" in scan_columns or "execution_model" in workspace_columns
752
763
  )
753
- if not (
754
- has_legacy_profile_history
755
- or has_public_warnings_history
756
- or has_legacy_profile_columns
757
- ):
764
+ if not (has_legacy_profile_history or has_legacy_profile_columns):
758
765
  return
759
766
 
760
767
  if (
@@ -796,17 +803,6 @@ def normalize_pre_release_execution_profile_migrations(
796
803
  "The Codex Security database has an unsupported execution-profile migration history."
797
804
  )
798
805
 
799
- if has_public_warnings_history:
800
- if execution_migrations.get(26) is not None:
801
- raise SystemExit(
802
- "The Codex Security database has an unsupported pre-release migration history."
803
- )
804
- connection.execute(
805
- "UPDATE schema_migrations SET version = 26 WHERE version = 25 AND name = ?",
806
- (warnings_migration_name,),
807
- )
808
- execution_migrations.pop(25)
809
-
810
806
  if execution_migrations.get(25) not in (None, model_migration_name):
811
807
  raise SystemExit(
812
808
  "The Codex Security database has an unsupported execution-profile migration history."
@@ -10,11 +10,10 @@ from typing import Any
10
10
 
11
11
  # Some plugin hosts launch Python with safe-path isolation enabled.
12
12
  sys.path.insert(0, str(Path(__file__).resolve().parent))
13
- from workbench_target import clean_worktree_content_digest, git_bytes, git_output
13
+ from workbench_target import clean_worktree_content_digest, git_bytes
14
14
 
15
15
  CONTEXT_LINES = 3
16
16
  MAX_BYTES = 16_000
17
- MAX_FILE_BYTES = 1_048_576
18
17
  MAX_LINES = 60
19
18
 
20
19
 
@@ -70,9 +69,6 @@ def scanned_source_text(scan: sqlite3.Row, target: Path, path: str) -> str | Non
70
69
  if snapshot_digest is not None and snapshot_digest != clean_worktree_content_digest():
71
70
  return None
72
71
  object_name = f"{revision}:{path}"
73
- size = git_output(target, "cat-file", "-s", object_name)
74
- if size is None or not size.isdigit() or int(size) > MAX_FILE_BYTES:
75
- return None
76
72
  content = git_bytes(target, "cat-file", "blob", object_name)
77
73
  return content.decode("utf-8", errors="replace") if content is not None else None
78
74