Graphene 3.4

Last reviewed Sep 20, 2026

Export as PDF

Prompt Details

Prompt Composition

2prompts
  1. Python 3155 lines · 3.8K tokens
  2. Graphene 3.467 lines · 3.4K tokens
Total222 lines · 7.2K tokens

CWE Coverage

  • 20
  • 23
  • 36
  • 59
  • 78
  • 89
  • 95
  • 117
  • 208
  • 209
  • 214
  • 250
  • 295
  • 306
  • 319
  • 328
  • 338
  • 347
  • 352
  • 367
  • 379
  • 409
  • 416
  • 426
  • 427
  • 470
  • 488
  • 502
  • 524
  • 532
  • 611
  • 639
  • 674
  • 732
  • 770
  • 772
  • 776
  • 807
  • 829
  • 862
  • 863
  • 915
  • 916
  • 918
  • 1104
  • 1333
  • 1395

Prompt Preview

Apply the Python 3 prompt first. These rules cover the Graphene 3.4 line with graphql-core 3.2 and were checked against Graphene 3.4.3. Graphene's tagged classifiers and test matrix cover Python 3.8–3.13; choose a maintained interpreter within that range and verify the complete stack before adopting a newer Python. Apply the chosen web framework and ORM adapter's controls separately: Graphene core supplies neither an HTTP/WebSocket server nor authentication, CSRF protection, or database transactions.

Version and schema boundary

  • Keep graphql-core on its 3.2 line, the newest line Graphene's declared >=3.1,<3.3 range allows and one graphql-core still patches, and refresh it in the lock file whenever you change Graphene, because that range also admits graphql-core releases with published parser and validation denial-of-service flaws. Keep the standard validation rules enabled, including OverlappingFieldsCanBeMergedRule, which the specification requires, and don't raise MAX_FIELD_COMPARISONS in graphql.validation.rules.overlapping_fields_can_be_merged. Keep graphql-relay within Graphene's declared >=3.1,<3.3 range.
  • On graphql-core 3.1.0 through 3.2.11, which predate the 3.2.12 fix for both flaws, overlapping-field validation has no comparison budget, and max_tokens, where it exists, doesn't count comments, so a highly compressible document a few kilobytes long on the wire can pin a worker for minutes or inflate memory. On those releases, cap the decompressed document at the size of your largest real operation before parse() and validate() run.