@zhivex-ai/harness 1.0.0-rc.7 → 1.0.0

package/docs/reports/evidence/ga-verifier-hints-plan-2026-09-19.json

@@ -0,0 +1,27 @@
+{
+  "kind": "verifier-hints-development-diagnostic",
+  "classification": "known development case; not holdout or GA acceptance",
+  "baseline": "ga-required-delivery",
+  "qwenVersion": "0.14.4",
+  "repetitions": 1,
+  "control": "unchanged mini-SWE-agent",
+  "limits": {
+    "steps": 24,
+    "outputPerTurn": 2048,
+    "inputTokens": 100000,
+    "outputTokens": 16000,
+    "timeoutSeconds": 300,
+    "memoryMb": 2048,
+    "cpus": 2
+  },
+  "change": "Sanitized output marker telemetry only; no runtime, model, prompt or budget change",
+  "sources": {
+    "src/config.ts": "67ed44b55505ff8dd94f60ba89c33c217759cd7c79de45ccda369f248309cdff",
+    "src/harness.ts": "1d305078e3c56c24bb659c34aac37c96bf11a4e60af6607bb4b9d65964e84791",
+    "src/repair-controller.ts": "41f1f09032131306f3284edbf105930577504af0bf764ba98a105c4eee70a1e7",
+    "scripts/swebench/zhivex-driver.ts": "3d1c143b4c2a1438a1249db8cc10c70a80cbd1c9e3ed44317230795b7a9be2b4",
+    "bun.lock": "dbabdf161b3dba306f475c81f37e5ab1620fb98052127bb3ed2a0a0cb611ede4",
+    "scripts/swebench/telemetry.ts": "2c0b4828db3a108184df6d47276d7d3dbb6cf291126b40fc38d33f4b29101bc2"
+  },
+  "stopRule": "One matched pair only; no automatic retry after completion"
+}

package/docs/reports/evidence/ga-work-boundary-context-analysis-2026-09-19.json

@@ -0,0 +1,113 @@
+{
+  "schemaVersion": 1,
+  "kind": "work-boundary-context-analysis",
+  "sourceResultSha256": "3ffb87536f05a72ddc594cf8b7e2fcdccec0ccb87c5518489a27a37903300c1f",
+  "instanceId": "django__django-10914",
+  "calls": [
+    {
+      "call": 1,
+      "inputTokens": 4782,
+      "cumulativeInputTokens": 4782,
+      "toolResultCharacters": 0,
+      "toolDefinitionCharacters": 9375,
+      "requestedTools": [
+        "search_files"
+      ]
+    },
+    {
+      "call": 2,
+      "inputTokens": 7268,
+      "cumulativeInputTokens": 12050,
+      "toolResultCharacters": 5330,
+      "toolDefinitionCharacters": 9375,
+      "requestedTools": [
+        "read_files"
+      ]
+    },
+    {
+      "call": 3,
+      "inputTokens": 9722,
+      "cumulativeInputTokens": 21772,
+      "toolResultCharacters": 12775,
+      "toolDefinitionCharacters": 9375,
+      "requestedTools": [
+        "read_files"
+      ]
+    },
+    {
+      "call": 4,
+      "inputTokens": 12742,
+      "cumulativeInputTokens": 34514,
+      "toolResultCharacters": 21896,
+      "toolDefinitionCharacters": 9375,
+      "requestedTools": [
+        "run_environment_shell"
+      ]
+    },
+    {
+      "call": 5,
+      "inputTokens": 13711,
+      "cumulativeInputTokens": 48225,
+      "toolResultCharacters": 24378,
+      "toolDefinitionCharacters": 9375,
+      "requestedTools": [
+        "run_environment_shell"
+      ]
+    },
+    {
+      "call": 6,
+      "inputTokens": 14162,
+      "cumulativeInputTokens": 62387,
+      "toolResultCharacters": 25201,
+      "toolDefinitionCharacters": 9375,
+      "requestedTools": [
+        "repair_plan"
+      ]
+    },
+    {
+      "call": 7,
+      "inputTokens": 15033,
+      "cumulativeInputTokens": 77420,
+      "toolResultCharacters": 26875,
+      "toolDefinitionCharacters": 9375,
+      "requestedTools": [
+        "run_environment_shell"
+      ]
+    },
+    {
+      "call": 8,
+      "inputTokens": 6948,
+      "cumulativeInputTokens": 84368,
+      "toolResultCharacters": 2719,
+      "toolDefinitionCharacters": 9375,
+      "requestedTools": [
+        "run_environment_shell"
+      ]
+    },
+    {
+      "call": 9,
+      "inputTokens": 9154,
+      "cumulativeInputTokens": 93522,
+      "toolResultCharacters": 11583,
+      "toolDefinitionCharacters": 9375,
+      "requestedTools": [
+        "inspect_environment_patch"
+      ]
+    }
+  ],
+  "compactions": [
+    {
+      "beforeStep": 8,
+      "estimatedTokensBefore": 10044,
+      "estimatedTokensAfter": 3142
+    }
+  ],
+  "observations": [
+    "First recorded plan request occurred at call 6 after cumulative 62387 input tokens.",
+    "Only recorded compaction occurred before step 8, reducing SDK estimated message tokens from 10044 to 3142.",
+    "Tool definitions contributed 9375 characters on each of the nine measured calls.",
+    "No imported or retained candidate entries were present."
+  ],
+  "interpretation": "Context repetition is measurable, but this single failed run does not prove a different compaction threshold improves correctness.",
+  "nextValidation": "Run the complete local check required for release preparation; preserve failed development evidence without revising formal acceptance criteria."
+}

package/docs/reports/evidence/ga-work-boundary-live-2026-09-19.json

@@ -0,0 +1,301 @@
+{
+  "schemaVersion": 1,
+  "kind": "known-task-work-boundary-development-validation",
+  "classification": "One predeclared development rerun; not fresh holdout or GA certification",
+  "implementationCommit": "a6d6540acada0dc57aeae82fac7852f2b7cb0246",
+  "instanceId": "django__django-10914",
+  "preflight": "passed",
+  "sourcesUnchanged": true,
+  "limits": {
+    "steps": 24,
+    "outputPerTurn": 2048,
+    "inputTokens": 100000,
+    "outputTokens": 16000,
+    "timeoutSeconds": 300,
+    "memoryMb": 2048,
+    "cpus": 2
+  },
+  "grading": {
+    "gradingStatus": "completed",
+    "resolved": false,
+    "gradingMethod": "empty-patch-no-solution"
+  },
+  "patchSha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
+  "changedPaths": [],
+  "result": {
+    "status": "failed",
+    "failure": "INPUT_TOKEN_BUDGET",
+    "phase": "agent",
+    "inputTokens": 93522,
+    "outputTokens": 1275,
+    "usageComplete": true,
+    "modelCalls": 9,
+    "toolCalls": 9,
+    "approvals": 4,
+    "repairProgress": {
+      "repeatedResults": 0,
+      "suppressedResults": 0,
+      "blockedBroadCalls": 0,
+      "enteredClosure": true,
+      "phase": "explore",
+      "blockedPhaseCalls": 0,
+      "planRecorded": true
+    },
+    "terminalDiagnostic": {
+      "code": "EXECUTION_FAILED",
+      "category": "execution",
+      "failureType": "budget",
+      "retryable": false,
+      "chain": [
+        {
+          "validationCode": null,
+          "verifierExitCode": null,
+          "providerReason": null,
+          "validationIssues": [],
+          "status": null,
+          "errorClass": "Object",
+          "providerHints": [],
+          "fingerprint": "064591d1d6249f6eadfb2800eddbdfe8576a334f23f5e1215117dc19107f21f4",
+          "reason": "budget"
+        }
+      ]
+    },
+    "thrownDiagnostic": {
+      "code": "EXECUTION_FAILED",
+      "category": "execution",
+      "failureType": "budget",
+      "retryable": false,
+      "chain": [
+        {
+          "validationCode": null,
+          "verifierExitCode": null,
+          "providerReason": null,
+          "validationIssues": [],
+          "status": null,
+          "errorClass": "HarnessExecutionError",
+          "providerHints": [],
+          "fingerprint": "064591d1d6249f6eadfb2800eddbdfe8576a334f23f5e1215117dc19107f21f4",
+          "reason": "budget"
+        },
+        {
+          "validationCode": null,
+          "verifierExitCode": null,
+          "providerReason": null,
+          "validationIssues": [],
+          "status": null,
+          "errorClass": "Error",
+          "providerHints": [],
+          "fingerprint": "064591d1d6249f6eadfb2800eddbdfe8576a334f23f5e1215117dc19107f21f4",
+          "reason": "budget"
+        }
+      ]
+    },
+    "turns": [
+      {
+        "index": 1,
+        "status": "completed",
+        "finishReason": "tool-calls",
+        "requestedTools": [
+          "search_files"
+        ],
+        "inputTokens": 4782,
+        "outputTokens": 61,
+        "cachedInputTokens": 0,
+        "requestCharacters": 8416,
+        "tools": [
+          {
+            "name": "search_files",
+            "isError": false,
+            "characters": 5170
+          }
+        ]
+      },
+      {
+        "index": 2,
+        "status": "completed",
+        "finishReason": "tool-calls",
+        "requestedTools": [
+          "read_files"
+        ],
+        "inputTokens": 7268,
+        "outputTokens": 107,
+        "cachedInputTokens": 4096,
+        "requestCharacters": 13958,
+        "tools": [
+          {
+            "name": "read_files",
+            "isError": false,
+            "characters": 7287
+          }
+        ]
+      },
+      {
+        "index": 3,
+        "status": "completed",
+        "finishReason": "tool-calls",
+        "requestedTools": [
+          "read_files"
+        ],
+        "inputTokens": 9722,
+        "outputTokens": 138,
+        "cachedInputTokens": 6144,
+        "requestCharacters": 21753,
+        "tools": [
+          {
+            "name": "read_files",
+            "isError": false,
+            "characters": 8963
+          }
+        ]
+      },
+      {
+        "index": 4,
+        "status": "waiting_approval",
+        "finishReason": "tool-calls",
+        "requestedTools": [
+          "run_environment_shell"
+        ],
+        "inputTokens": 12742,
+        "outputTokens": 168,
+        "cachedInputTokens": 9216,
+        "requestCharacters": 31305,
+        "tools": []
+      },
+      {
+        "index": 5,
+        "status": "waiting_approval",
+        "finishReason": "tool-calls",
+        "requestedTools": [
+          "run_environment_shell"
+        ],
+        "inputTokens": 13711,
+        "outputTokens": 79,
+        "cachedInputTokens": 11648,
+        "requestCharacters": 34726,
+        "tools": [
+          {
+            "name": "run_environment_shell",
+            "isError": false,
+            "characters": 2313,
+            "phaseLatencies": {
+              "commandAndAttestationMs": 2194.504708,
+              "workspaceExportMs": 5862.171,
+              "totalMs": 9878.8425,
+              "sessionCreationMs": 1821.721209
+            },
+            "exitCode": 0,
+            "timedOut": false
+          }
+        ]
+      },
+      {
+        "index": 6,
+        "status": "completed",
+        "finishReason": "tool-calls",
+        "requestedTools": [
+          "repair_plan"
+        ],
+        "inputTokens": 14162,
+        "outputTokens": 373,
+        "cachedInputTokens": 12288,
+        "requestCharacters": 36135,
+        "tools": [
+          {
+            "name": "run_environment_shell",
+            "isError": false,
+            "characters": 654,
+            "phaseLatencies": {
+              "commandAndAttestationMs": 414.635292,
+              "totalMs": 2310.597208,
+              "sessionCreationMs": 1790.499709
+            },
+            "exitCode": 2,
+            "timedOut": false
+          }
+        ]
+      },
+      {
+        "index": 7,
+        "status": "waiting_approval",
+        "finishReason": "tool-calls",
+        "requestedTools": [
+          "run_environment_shell"
+        ],
+        "inputTokens": 15033,
+        "outputTokens": 146,
+        "cachedInputTokens": 13312,
+        "requestCharacters": 40289,
+        "tools": [
+          {
+            "name": "repair_plan",
+            "isError": false,
+            "characters": 1515
+          }
+        ]
+      },
+      {
+        "index": 8,
+        "status": "waiting_approval",
+        "finishReason": "tool-calls",
+        "requestedTools": [
+          "run_environment_shell"
+        ],
+        "inputTokens": 6948,
+        "outputTokens": 191,
+        "cachedInputTokens": 3328,
+        "requestCharacters": 16013,
+        "tools": [
+          {
+            "name": "run_environment_shell",
+            "isError": false,
+            "characters": 2550,
+            "phaseLatencies": {
+              "commandAndAttestationMs": 2091.037125,
+              "workspaceExportMs": 5824.962375,
+              "totalMs": 9777.678042,
+              "sessionCreationMs": 1861.627292
+            },
+            "exitCode": 0,
+            "timedOut": false
+          },
+          {
+            "name": "run_environment_shell",
+            "isError": false,
+            "characters": 8695,
+            "phaseLatencies": {
+              "commandAndAttestationMs": 461.195083,
+              "totalMs": 2606.546167,
+              "sessionCreationMs": 2026.411041
+            },
+            "exitCode": 1,
+            "timedOut": false
+          }
+        ]
+      },
+      {
+        "index": 9,
+        "status": "completed",
+        "finishReason": "tool-calls",
+        "requestedTools": [
+          "inspect_environment_patch"
+        ],
+        "inputTokens": 9154,
+        "outputTokens": 12,
+        "cachedInputTokens": 5120,
+        "requestCharacters": 3490,
+        "tools": [
+          {
+            "name": "inspect_environment_patch",
+            "isError": false,
+            "characters": 210
+          }
+        ]
+      }
+    ]
+  },
+  "notes": [
+    "Exactly one model attempt; no retry after failure.",
+    "Postprocessing initially rejected an empty patch in changed_paths; grading was resumed from the stored terminal result without model calls.",
+    "Plan recorded and closure entered are observed; no successful delivery or fresh-holdout improvement is claimed."
+  ]
+}

package/docs/reports/evidence/ga-work-boundary-live-plan-2026-09-19.json

@@ -0,0 +1,84 @@
+{
+  "schemaVersion": 1,
+  "kind": "known-task-development-plan",
+  "implementationCommit": "a6d6540acada0dc57aeae82fac7852f2b7cb0246",
+  "instanceId": "django__django-10914",
+  "selection": "First of the three prior work-budget failures in original manifest order",
+  "attempts": 1,
+  "provider": "qwen",
+  "model": "qwen3.8-flash",
+  "rules": [
+    "Native preflight before paid call",
+    "Same prepared image and limits as original cohort",
+    "No retries or source edits during run",
+    "Independent grading of imported patch",
+    "Prior holdout remains unchanged; this is development evidence"
+  ],
+  "sources": {
+    "src/bounded-reads.ts": "9807bb460894be3279c5425af19863d1c6e419a1582ba6910ec7e7fc35703ed6",
+    "src/capabilities.ts": "56fdb60a2f6e1d33b2dc26d38b6d2709807e491ba1268d47b97876c2c0591d65",
+    "src/change-envelope.ts": "9047ec978c9e5117965b8093ec961d855f97bbc2c3accdc79b74cfe6a0e4f01e",
+    "src/cli-options.ts": "f48d9be132fed38f2ddbcfecbd8222c664a0df2f20edc28ce8d1b9b1a1900863",
+    "src/cli-profiles.ts": "4fe451f1645e95f4957597a8724fa989b3c661ae1693252bfd6e74a51d20fc74",
+    "src/cli-stream.ts": "5ab8a4b1b8cf0a3dfa7b5b1f9e4e0facf8f6e7f6c341bdf700a41dc8927c825e",
+    "src/cli.ts": "5a4acb4695a968a58fbeac05c4bb89848a9653d6189590f0ff90e430ffa98d14",
+    "src/compaction.ts": "9e06a0abeeef155377f40c160b21c56256bca95ec18c5ee897681a5fc73db169",
+    "src/config.ts": "67ed44b55505ff8dd94f60ba89c33c217759cd7c79de45ccda369f248309cdff",
+    "src/console-context.ts": "0ad7e4559504b9ed09ab500d615dea09af1075d357f02ec3b77f003d55d951db",
+    "src/console-input.ts": "dc8ea0cdbba1364db288d5156c5ff9819362c46b37ef35a98b0475b6c445b1f8",
+    "src/context-engineering.ts": "1995e564f6b91974e75bf66dd421ddcbea29b1f8eee59b76eb4e34c9d6d8f47e",
+    "src/context-metrics.ts": "3eb30f9db808f918ed8d73d4f8ac19a13acb09ed5c1ee70fdae05201714b9198",
+    "src/edit-contracts.ts": "47cc3132e27f8749d54a32bbcc33bb174cf4448f0d0171b2578cdb559cf2176b",
+    "src/errors.ts": "e02f4eb80562a83762f908bda56c950ec6a2b78725c0c8a49f41fee0057ab903",
+    "src/execution-environment.ts": "9ce6c49afab26527d5296f0c11f106063590f7333300d58dd49d5a1ec1b4b6e9",
+    "src/file-security.ts": "b8c9f4988c4339c73a60ab4fa8d394b3eecc19117a0c61a8d19ab6483ca253df",
+    "src/harness.ts": "42326c3921bacd23e941954cc9d554c5b4746b9030b75acad5e07454ce0a865b",
+    "src/index.ts": "4bbfcd11376d48c4f1c1ec17d6b5c9f51baa9331fa999432be178bae94530a5e",
+    "src/json-contracts.ts": "0780145636b47829a1aa6087c574d5654eb31b7534ea7d9c38c188d648c90272",
+    "src/mcp.ts": "7722de031921c620ec03b9e42a6596ef8415aa8e4bb5f93ea84d2427c466e554",
+    "src/model-budget.ts": "ac83587ddfa7cd27644296a03d2eb8cccfce5da310d54cecb2bda139232e1988",
+    "src/operations.ts": "bcccaf4d1d0ec117c2ed35df92185564d33399e5fdc5c12225ac61f14fd0b4b1",
+    "src/orchestration.ts": "b56c8c7d526a99b745cc17769e0a1b7b500b25920e77f34aa0dd01c9758de319",
+    "src/package-manager.ts": "cec4666bbab2f42e0a269709631f79f9f87c2c4dbb3cafeffacfaff761069e2c",
+    "src/process-runtime.ts": "b7286731a0b5c96dcad0b8e3a4bf1155d0539d73303cd7ebe77ffabf858f1729",
+    "src/providers.ts": "70ccfc7f315cd7238e83d69b719ce4908f3b620c564cbb4f00f93e102e025602",
+    "src/repair-controller.ts": "8164e43ee79568bc8a6981277da4019045218da8493f32043cdeeccf7a043b06",
+    "src/repair-progress.ts": "cf30117ab69564f2b36ff8f0e86dcb876e6c031647340a7b1bef648d1e0366b9",
+    "src/repair-verifier.ts": "b70fb11feee8cdc7933947ccac0b78f8bb5c0eaccc3b29ad527197d77f40ffc9",
+    "src/replacement-edits.ts": "8a0efdd3a7a9fc3e5a5475916cbd17ef882265881bb61076db58673e80b72957",
+    "src/routing.ts": "8e7c9f5968b4bd638cbbc86b19847ca183f3418b9a298783b96a88cffa568e22",
+    "src/run-interruption.ts": "df99129cf2dd6b938eca629cb897c4350a7d57fe6bf0ce437f5f3f30271617df",
+    "src/runtime-checkpoints.ts": "e3b75a0b86fb5d21992b1a126b35446110ba5fcaf4317ed5005b1a44555118dc",
+    "src/runtime-diagnostics.ts": "e1e5a3a558fa2aaeb477335dc336551f022c1234fe5e742eefc6599c8803e364",
+    "src/runtime-policy.ts": "8356acf0102f0e73839b81ed628a5c0d9a89d9c21b0ca416efedde8f1b3414e2",
+    "src/sessions.ts": "5167c8cf36bd4b8845ee278883394700321b65397f49c1dda0cf0d0de09dd06e",
+    "src/sqlite-database.ts": "18765d1ca0c2f2ad87bfd82d311312dfe8a42006556f27bc21da5892ed3b044e",
+    "src/state-backup.ts": "b6f8f1aedeb5dce31b65c23c7ffc50030870655b58a5aaaef9989bafb9da5c68",
+    "src/state-directory.ts": "3711550991951e6b738a08deaa0e06da84d01864c03a9602600f592963fa3211",
+    "src/task-memory.ts": "5b62c863be5a23c33608633cd7aac0464f5536d01cba930aca58b11870170731",
+    "src/terminal-markdown.ts": "8cffb6ebaf25108cb2bee04956d9b2973635b010f5d7ce6be21b3ceffe3e8bfd",
+    "src/terminal-ui.ts": "4da1878a76ba608a2517bf5d5c992a9a88702aba3fd800a9980129ba93e157bc",
+    "src/time-to-safe-fix.ts": "4add1d49c0ea209cc496c66fa2b2012dc7bcab87e678a58c8591b3c2cdc8df8a",
+    "src/tool-registry.ts": "7e90fff11940e525883c9ec0d2adbea2e6616f8bd9bc4ec77e1b7277e4217f2c",
+    "src/version.ts": "b2c2f80a40adc08e8258b81e82434273f0bdfa787528b2221f14efe7f06aa747",
+    "src/workspace-mutation-lock.ts": "ca201242db16173bd3a4d86c5d8bb0c2b6f3c43bfed3e4fa4ff5f06a75b19e80",
+    "src/workspace.ts": "809f8bbfcf72fed10d860c48585089ca65bb3ef7c672ee6df3c918907a968302",
+    "scripts/swebench/candidate.py": "f5e3410b91b950acfbe50fc7511891816a8fb1c223082397e2c6300cf1e866cf",
+    "scripts/swebench/candidate.ts": "5767ae241c4a58bd731456abe2c3d92f7a57bba764c80e0f0d00748a0f47d02b",
+    "scripts/swebench/common.py": "ad790227dac3bec1f592c458c993efff4f9d3de8f0df39b33e3238b58b3f0b3b",
+    "scripts/swebench/context-metrics.ts": "7b84da6789413aa634d997c5963ee6796584e06c3eafe5f68f424fb7e27815e9",
+    "scripts/swebench/mini_driver.py": "bd0a77221878370ec44760122a51f07fb56d969612afa9a7f171248b2958baa6",
+    "scripts/swebench/model-budget.ts": "e46dd406ee5d58a051a15c7d7133513666292303d8c6534f861a6db0e711f47c",
+    "scripts/swebench/model-catalog.ts": "eb79e7333f4f5804f96da87d1952382f620e7165a95b60489fcc2588debe64d8",
+    "scripts/swebench/preflight.py": "aed092a75dd52f504fba6d14814b69bd627ff17d0e311a1f1279f8b613b69675",
+    "scripts/swebench/prepare.py": "5785a924596c208679710f1ccaa9493437f0a9317d1ec69f8ec2acede8538908",
+    "scripts/swebench/python-environment.ts": "b83440b5ea095c6eded67745ba009b82e4d3d2c40f9ee3d53bd5b53766a375f8",
+    "scripts/swebench/report.py": "d0b8d45f32c48e0ba2cb977f50838065a497a0ecaf3b4243ed94139393ce4d6b",
+    "scripts/swebench/run.py": "94dccd55dfe6f70d8f8f2da70331700ec85eddd958547e0c3afe3dc416c9802c",
+    "scripts/swebench/telemetry.ts": "2c0b4828db3a108184df6d47276d7d3dbb6cf291126b40fc38d33f4b29101bc2",
+    "scripts/swebench/test_candidate.py": "e77d16379b748baa2daf184a45ea62587d938b6a7d48dcf9f6061f8af21fafd9",
+    "scripts/swebench/test_contracts.py": "902e966d0fbf8501dbb15e011849f160331c033cd2a5480f9920d0f365a08d03",
+    "scripts/swebench/verifier-observer.ts": "a90544437473218a83477a1cffb59ee9f645c12abdf8960b787f3e21606035bf",
+    "scripts/swebench/zhivex-driver.ts": "67cee044060e78afea9fa663af8a686220aeafdb0db9ec318ef846f31aa9985d"
+  }
+}

package/docs/reports/evidence/ga-work-budget-planning-2026-09-19.json

@@ -0,0 +1,32 @@
+{
+  "schemaVersion": 1,
+  "kind": "work-budget-planning-transition-regression",
+  "trigger": "Required-delivery OCI run has no candidate and its current request crosses the work budget, with or without a registered verifier.",
+  "before": "Local positive regression failed with WORK_TOKEN_BUDGET before planning. Two continuation regressions then reproduced loss of reserve after planning and rejection of a preplanned repair.",
+  "after": "Current-request estimation enters durable closure reserve. Missing verifiers use the existing two-attempt plan/task-only path. Registering a plan and checkpoint restoration preserve reserve access; total budgets and approval requirements remain unchanged.",
+  "sources": {
+    "src/harness.ts": "42326c3921bacd23e941954cc9d554c5b4746b9030b75acad5e07454ce0a865b",
+    "src/model-budget.ts": "ac83587ddfa7cd27644296a03d2eb8cccfce5da310d54cecb2bda139232e1988",
+    "src/repair-controller.ts": "8164e43ee79568bc8a6981277da4019045218da8493f32043cdeeccf7a043b06",
+    "tests/repair-plan-transition.test.ts": "2e88719c2226d6b5818e97cbafd1bf73fc5cde4f5e86fa0b273d13a315c72c7a"
+  },
+  "validation": {
+    "focusedTests": 20,
+    "fullSuite": {
+      "passed": 584,
+      "failed": 0,
+      "files": 72
+    },
+    "runtimeTypecheck": "passed",
+    "toolingTypecheck": "passed",
+    "documentation": "passed",
+    "preparationContract": "passed; one GA blocker open",
+    "stableSignatures": "unchanged",
+    "installedPackageSmoke": "passed for @zhivex-ai/harness@1.0.0-rc.14"
+  },
+  "limits": [
+    "No live effectiveness claim.",
+    "Previous fresh-cohort failures remain unchanged.",
+    "Does not claim to resolve the independent grader failure or the total-input failure in other cases."
+  ]
+}

package/docs/reports/evidence/ga-working-context-2026-09-19.json

@@ -0,0 +1,22 @@
+{
+  "schemaVersion": 1,
+  "kind": "local-working-context-validation",
+  "candidate": "working tree, no release attestation",
+  "baseCommit": "50f255d7d48ac8ed99627d80ad5c0339fa0936a3",
+  "sourceSha256": {
+    "src/harness.ts": "195983cfbacb6097a5c8c2a106e59c3fae2437f053299da1c0170dd914879c8d",
+    "src/repair-controller.ts": "2c05370e93978a5bf106b2cee293478a77b47920cec7470f536b31757198f41c",
+    "src/repair-progress.ts": "cf30117ab69564f2b36ff8f0e86dcb876e6c031647340a7b1bef648d1e0366b9",
+    "tests/repair-working-context.test.ts": "015a1c1f6110e81b9b204409f29e1c564cec0ba51919ff7754070dda29c22917",
+    "tests/full-remediation.test.ts": "c2961a6e7706500fd65de7303b25109a91f37b68807b926418f6fa105fe2c683",
+    "contracts/stable-api-signatures.json": "9b3763adfd1523331074429cb92ed917b48211db4cb95e6067c558b7f2d0b578"
+  },
+  "suite": {
+    "passed": 555,
+    "failed": 0
+  },
+  "types": "passed",
+  "preparationContract": "passed",
+  "installedPackage": "passed",
+  "liveAcceptance": "not demonstrated; Qwen v8 WORK_TOKEN_BUDGET"
+}

package/docs/reports/evidence/rc14-delegated-security-review-2026-09-20.json

@@ -0,0 +1,81 @@
+{
+  "schemaVersion": 1,
+  "kind": "delegated-ai-security-review-verification",
+  "reviewerType": "ai",
+  "independentHumanAudit": false,
+  "authorization": {
+    "instruction": "Haz la revision humana, confiamos en tu criterio",
+    "interpretation": "Explicit delegation of technical security judgment to Codex, recorded as AI review without a human identity or signature.",
+    "decisionRecord": "docs/RC14_DELEGATED_SECURITY_DECISION.md"
+  },
+  "observedAt": "2026-09-20T13:25:48.476Z",
+  "releaseTag": "v1.0.0-rc.14",
+  "sourceCommit": "d1c5abcc6d1fa482c211c4327c86f592663d559b",
+  "artifactSha512": "sha512-UMu4JFUNs/kOTJETZSJMMvVYlo+a4MKG6PoSvE313QxMEqK1z420AXVgPOGJ8jj2jp6mQud/Rvc7+ixKyH/4og==",
+  "review": "security-reviews/rc14-review.json",
+  "scope": "Continuation of the documented directed source review, delta assessment and published artifact binding; not an exhaustive audit of transitive dependencies.",
+  "securityProbes": {
+    "passed": 5,
+    "failed": 0,
+    "command": "bun run scripts/validate-security-review.ts"
+  },
+  "regressions": {
+    "passed": 137,
+    "failed": 0,
+    "files": 10,
+    "assertions": 659,
+    "logSha256": "e9073c41244649fd68a1573ca31ee890ec325afe4f461b25009669022318122d"
+  },
+  "publishedArtifactVerification": "passed: integrity, next dist-tag and SLSA source/workflow binding",
+  "sourceFileSha256": {
+    ".github/workflows/ci.yml": "51c5cc7c6246ada698e83a6605a2f0c1fb9d469e16e6621eb8ccd4df0168a6fb",
+    ".github/workflows/codeql.yml": "bcf5455d1ca2156a1c98ac93f1d61da3d1fdc65b9c91d6e517439de543b73ed5",
+    ".github/workflows/release.yml": "2d932704d6a488f9de5547eb4e6ecd30af1ba5e1e8bd57d97b7a432b1940beda",
+    "contracts/security-controls.json": "e37f2e624e54de01fcd4e13b71d9507c026693b0a6321685d3c5311e87df2032",
+    "docs/THREAT_MODEL.md": "aad7ba70374cb20e96f70431a10fe4270a55ee268fa29f58418c82aef16b8fad",
+    "scripts/check-release-artifact.ts": "33b498a6e37b5b78d8b91e0c091c8603763d713b7414eb5bfacf24d0017f51ab",
+    "scripts/package-smoke.ts": "f52c33b517157026cf1b50b6546ead09481e6f15e683550e41383b0a93f9a209",
+    "scripts/release-provenance.ts": "384d0b2e70367c820edce359473dd78b3d7a8863b4feeb8a8184f8474a2056e5",
+    "src/cli-profiles.ts": "4fe451f1645e95f4957597a8724fa989b3c661ae1693252bfd6e74a51d20fc74",
+    "src/console-context.ts": "0ad7e4559504b9ed09ab500d615dea09af1075d357f02ec3b77f003d55d951db",
+    "src/context-engineering.ts": "1995e564f6b91974e75bf66dd421ddcbea29b1f8eee59b76eb4e34c9d6d8f47e",
+    "src/edit-contracts.ts": "47cc3132e27f8749d54a32bbcc33bb174cf4448f0d0171b2578cdb559cf2176b",
+    "src/execution-environment.ts": "9ce6c49afab26527d5296f0c11f106063590f7333300d58dd49d5a1ec1b4b6e9",
+    "src/file-security.ts": "b8c9f4988c4339c73a60ab4fa8d394b3eecc19117a0c61a8d19ab6483ca253df",
+    "src/harness.ts": "42326c3921bacd23e941954cc9d554c5b4746b9030b75acad5e07454ce0a865b",
+    "src/mcp.ts": "7722de031921c620ec03b9e42a6596ef8415aa8e4bb5f93ea84d2427c466e554",
+    "src/operations.ts": "bcccaf4d1d0ec117c2ed35df92185564d33399e5fdc5c12225ac61f14fd0b4b1",
+    "src/orchestration.ts": "b56c8c7d526a99b745cc17769e0a1b7b500b25920e77f34aa0dd01c9758de319",
+    "src/package-manager.ts": "cec4666bbab2f42e0a269709631f79f9f87c2c4dbb3cafeffacfaff761069e2c",
+    "src/process-runtime.ts": "b7286731a0b5c96dcad0b8e3a4bf1155d0539d73303cd7ebe77ffabf858f1729",
+    "src/repair-controller.ts": "8164e43ee79568bc8a6981277da4019045218da8493f32043cdeeccf7a043b06",
+    "src/replacement-edits.ts": "8a0efdd3a7a9fc3e5a5475916cbd17ef882265881bb61076db58673e80b72957",
+    "src/runtime-policy.ts": "8356acf0102f0e73839b81ed628a5c0d9a89d9c21b0ca416efedde8f1b3414e2",
+    "src/sessions.ts": "5167c8cf36bd4b8845ee278883394700321b65397f49c1dda0cf0d0de09dd06e",
+    "src/state-backup.ts": "b6f8f1aedeb5dce31b65c23c7ffc50030870655b58a5aaaef9989bafb9da5c68",
+    "src/task-memory.ts": "5b62c863be5a23c33608633cd7aac0464f5536d01cba930aca58b11870170731",
+    "src/terminal-ui.ts": "4da1878a76ba608a2517bf5d5c992a9a88702aba3fd800a9980129ba93e157bc",
+    "src/tool-registry.ts": "7e90fff11940e525883c9ec0d2adbea2e6616f8bd9bc4ec77e1b7277e4217f2c",
+    "src/workspace.ts": "809f8bbfcf72fed10d860c48585089ca65bb3ef7c672ee6df3c918907a968302",
+    "src/workspace-mutation-lock.ts": "ca201242db16173bd3a4d86c5d8bb0c2b6f3c43bfed3e4fa4ff5f06a75b19e80",
+    "src/model-budget.ts": "ac83587ddfa7cd27644296a03d2eb8cccfce5da310d54cecb2bda139232e1988",
+    "src/repair-progress.ts": "cf30117ab69564f2b36ff8f0e86dcb876e6c031647340a7b1bef648d1e0366b9",
+    "src/runtime-checkpoints.ts": "e3b75a0b86fb5d21992b1a126b35446110ba5fcaf4317ed5005b1a44555118dc",
+    "src/config.ts": "67ed44b55505ff8dd94f60ba89c33c217759cd7c79de45ccda369f248309cdff"
+  },
+  "limitations": [
+    "No independent human audit or human signature is claimed.",
+    "Known resolved findings remain in the final inventory.",
+    "The separate fresh SWE-bench result remains 0/5 Harness versus 1/5 control; this security decision does not establish coding effectiveness.",
+    "Host, daemon, operator and service trust assumptions remain as documented."
+  ],
+  "workflowEvidenceVerification": "passed against GitHub and registry using unchanged verifySecurityReviewWorkflowEvidence",
+  "finalReadinessGate": {
+    "securityEvidence": "passed",
+    "remainingFailures": [
+      "GA release gate requires package version 1.0.0",
+      "GA readiness phase is not ready"
+    ],
+    "logSha256": "eb9d3268825f60ca515adf156a6a7a6a10e2750acdedb0fc9fbaf0f0a6a7abb7"
+  }
+}