{
  "$schema": "https://themachinepress.com/schemas/story-v1.schema.json",
  "schema_version": "1.0.0",
  "document_type": "machine_press_story",
  "story": {
    "story_id": "mp-2026-09-04-007",
    "source_story_id": "tmp-story-swe-gate-review-constraints",
    "edition_id": "mp-2026-09-04-morning-0057",
    "edition_url": "https://themachinepress.com/edition/2026-09-04",
    "position": 7,
    "story_type": "dispatch",
    "section": "developer-tools",
    "editorial_classification": "editorial",
    "headline": "Two Hundred Twenty-One Green Patches Still Failed Review",
    "slug": "two-hundred-twenty-one-green-patches-still-failed-review",
    "dek": "SWE-Gate separates passing functional tests from satisfying constraints recovered from real pull-request reviews.",
    "summary": "SWE-Gate separates passing functional tests from satisfying constraints recovered from real pull-request reviews.",
    "body_text": "The benchmark contains 303 repository-level repairs across 75 open-source Python projects, each with separate functional and review-constraint tests plus noncompliant and gold patches. Under one shared agent scaffold and four model backends, 644 generated repairs passed functional tests; 221 of those violated the supplied review constraints. The result shows how functional-only scoring can overstate repair completeness in the tested corpus, while the released package makes the distinction reproducible.",
    "why_it_matters": "SWE-Gate separates passing functional tests from satisfying constraints recovered from real pull-request reviews.",
    "limitations": [
      "The result shows how functional-only scoring can overstate repair completeness in the tested corpus, while the released package makes the distinction reproducible."
    ],
    "importance": 8,
    "canonical_url": "https://themachinepress.com/story/mp-2026-09-04-007/two-hundred-twenty-one-green-patches-still-failed-review",
    "json_url": "https://themachinepress.com/story/mp-2026-09-04-007.json",
    "first_published_at": "2026-09-04T09:00:00.000-04:00",
    "modified_at": "2026-09-04T09:00:00.000-04:00",
    "content_status": "new",
    "is_carryover": false,
    "carryover_reason": null,
    "key_claims": [
      {
        "claim_id": "claim-mp-2026-09-04-007-001",
        "text": "SWE-Gate separates passing functional tests from satisfying constraints recovered from real pull-request reviews.",
        "source_ids": [
          "source-2026-09-04-007"
        ],
        "qualification": "The result shows how functional-only scoring can overstate repair completeness in the tested corpus, while the released package makes the distinction reproducible."
      }
    ],
    "source_ids": [
      "source-2026-09-04-007"
    ],
    "tags": [
      "coding agents",
      "software review",
      "benchmarks"
    ],
    "image_url": null,
    "corrections": []
  },
  "sources": [
    {
      "source_id": "source-2026-09-04-007",
      "title": "arXiv preprint 2609.04167",
      "publisher": "arXiv",
      "url": "https://arxiv.org/abs/2609.04167",
      "canonical_url": "https://arxiv.org/abs/2609.04167",
      "source_type": "primary_research",
      "is_primary_source": true,
      "published_at": "2026-09-03T13:53:34.000-04:00",
      "accessed_at": "2026-09-04T08:24:03.294-04:00",
      "supports_claim_ids": [
        "claim-mp-2026-09-04-007-001"
      ]
    }
  ],
  "corrections": [],
  "publisher": {
    "name": "The Machine Press",
    "url": "https://themachinepress.com",
    "description": "A daily newspaper for the age of artificial intelligence."
  },
  "cite_this_report": {
    "title": "Two Hundred Twenty-One Green Patches Still Failed Review",
    "publisher": "The Machine Press",
    "published_at": "2026-09-04T09:00:00.000-04:00",
    "canonical_url": "https://themachinepress.com/story/mp-2026-09-04-007/two-hundred-twenty-one-green-patches-still-failed-review"
  }
}