@piagent/platform 1.2.17 → 1.3.0-ie.8

package/evals/fixtures/runtime-model-snapshot.invalid.json

@@ -0,0 +1,18 @@
+{
+  "schemaVersion": 1,
+  "capturedAt": "tomorrow",
+  "source": "api-doc-hint",
+  "piHostVersion": null,
+  "piagentVersion": null,
+  "provider": "openai",
+  "modelId": "example",
+  "availability": "public",
+  "contextWindow": -1,
+  "requestedThinkingLevel": null,
+  "effectiveThinkingLevel": null,
+  "supportedThinkingLevels": [],
+  "capabilities": [],
+  "provenance": [],
+  "warnings": [],
+  "accessToken": "synthetic-secret"
+}

package/evals/fixtures/runtime-model-snapshot.valid.json

@@ -0,0 +1,21 @@
+{
+  "schemaVersion": 1,
+  "capturedAt": "2026-08-08T00:00:00.000Z",
+  "source": "pi-runtime",
+  "piHostVersion": "0.82.0",
+  "piagentVersion": "1.2.17",
+  "provider": "openai-codex",
+  "modelId": "gpt-5.6-terra",
+  "availability": "authenticated",
+  "contextWindow": 200000,
+  "requestedThinkingLevel": "medium",
+  "effectiveThinkingLevel": "medium",
+  "supportedThinkingLevels": null,
+  "capabilities": [
+    { "name": "image-input", "value": null, "source": "pi-runtime" }
+  ],
+  "provenance": [
+    { "field": "modelId", "source": "pi-runtime", "capturedAt": "2026-08-08T00:00:00.000Z" }
+  ],
+  "warnings": []
+}

package/evals/fixtures/solver-decision.invalid.json

@@ -0,0 +1 @@
+{"schemaVersion":1,"policyVersion":"solver-v1","featureHash":"aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa","route":"implement-everything","plannedPhases":[],"context":{"recommendation":"none","budgetBand":"none"},"toolGroups":[],"helper":{"needed":true,"role":"worker","enforced":true},"parentModel":{"provider":null,"modelId":null,"effort":null,"enforced":true},"reasonCodes":[],"confidence":"certain","mode":"shadow","override":{"observed":false,"route":null,"recordedAt":null}}

package/evals/fixtures/solver-decision.valid.json

@@ -0,0 +1 @@
+{"schemaVersion":1,"policyVersion":"solver-v1","featureHash":"aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa","route":"direct","plannedPhases":["intake","implement","verify"],"context":{"recommendation":"targeted","budgetBand":"small"},"toolGroups":["intake","task"],"helper":{"needed":false,"role":null,"enforced":false},"parentModel":{"provider":"openai-codex","modelId":"gpt-5.6-terra","effort":"medium","enforced":false},"reasonCodes":["explicit-path"],"confidence":"high","mode":"shadow","override":{"observed":false,"route":null,"recordedAt":null}}

package/evals/fixtures/task-authority-snapshot.invalid.json

@@ -0,0 +1,6 @@
+{
+  "schemaVersion": 2,
+  "snapshotVersion": "unknown-task-authority-snapshot",
+  "manifestVersion": "unknown-authority-version",
+  "manifestDigest": "sha256:ffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff"
+}

package/evals/fixtures/task-authority-snapshot.valid.json

@@ -0,0 +1,279 @@
+{
+  "schemaVersion": 1,
+  "snapshotVersion": "task-authority-snapshot-v1",
+  "manifestVersion": "authority-v1",
+  "manifestDigest": "sha256:28112cb62a8868d4371f78c9edefff7489adb0dc6616b9e489c4303b80307393",
+  "releaseVersion": "1.3.0",
+  "profile": "broad-default",
+  "taskId": "task-authority-fixture",
+  "taskRunId": "run-authority-fixture",
+  "capturedAt": "2026-08-10T00:00:00.000Z",
+  "resolution": {
+    "source": "profile",
+    "modeOverrides": {}
+  },
+  "globalBudgets": {
+    "maxSystemContinuationsPerTask": 1,
+    "maxAutomaticDispatchesPerTask": 1,
+    "maxSpecialistReviewRoundsPerTask": 1
+  },
+  "capabilities": [
+    {
+      "id": "CAP-01",
+      "owner": "task-contract-maintainer",
+      "mode": "enforce",
+      "authority": "enforce",
+      "dependencies": [],
+      "budgets": {
+        "systemContinuations": 0,
+        "automaticDispatches": 0,
+        "reviewRounds": 0
+      }
+    },
+    {
+      "id": "CAP-02",
+      "owner": "security-policy-maintainer",
+      "mode": "enforce",
+      "authority": "enforce",
+      "dependencies": [
+        "CAP-01"
+      ],
+      "budgets": {
+        "systemContinuations": 0,
+        "automaticDispatches": 0,
+        "reviewRounds": 0
+      }
+    },
+    {
+      "id": "CAP-03",
+      "owner": "verification-truth-maintainer",
+      "mode": "enforce",
+      "authority": "enforce",
+      "dependencies": [
+        "CAP-01",
+        "CAP-02"
+      ],
+      "budgets": {
+        "systemContinuations": 0,
+        "automaticDispatches": 0,
+        "reviewRounds": 0
+      }
+    },
+    {
+      "id": "CAP-04",
+      "owner": "long-task-state-maintainer",
+      "mode": "enforce",
+      "authority": "enforce",
+      "dependencies": [
+        "CAP-01",
+        "CAP-03"
+      ],
+      "budgets": {
+        "systemContinuations": 0,
+        "automaticDispatches": 0,
+        "reviewRounds": 0
+      }
+    },
+    {
+      "id": "CAP-05",
+      "owner": "context-efficiency-maintainer",
+      "mode": "on",
+      "authority": "enforce",
+      "dependencies": [
+        "CAP-01"
+      ],
+      "budgets": {
+        "systemContinuations": 0,
+        "automaticDispatches": 0,
+        "reviewRounds": 0
+      }
+    },
+    {
+      "id": "CAP-06",
+      "owner": "tool-surface-maintainer",
+      "mode": "on",
+      "authority": "enforce",
+      "dependencies": [
+        "CAP-02"
+      ],
+      "budgets": {
+        "systemContinuations": 0,
+        "automaticDispatches": 0,
+        "reviewRounds": 0
+      }
+    },
+    {
+      "id": "CAP-07",
+      "owner": "retrieval-memory-maintainer",
+      "mode": "recommend",
+      "authority": "advise",
+      "dependencies": [
+        "CAP-02",
+        "CAP-05"
+      ],
+      "budgets": {
+        "systemContinuations": 0,
+        "automaticDispatches": 0,
+        "reviewRounds": 0
+      }
+    },
+    {
+      "id": "CAP-08",
+      "owner": "solver-runtime-maintainer",
+      "mode": "shadow",
+      "authority": "observe",
+      "dependencies": [
+        "CAP-01",
+        "CAP-05"
+      ],
+      "budgets": {
+        "systemContinuations": 0,
+        "automaticDispatches": 0,
+        "reviewRounds": 0
+      }
+    },
+    {
+      "id": "CAP-09",
+      "owner": "phase-policy-maintainer",
+      "mode": "shadow",
+      "authority": "observe",
+      "dependencies": [
+        "CAP-01",
+        "CAP-02",
+        "CAP-10"
+      ],
+      "budgets": {
+        "systemContinuations": 0,
+        "automaticDispatches": 0,
+        "reviewRounds": 0
+      }
+    },
+    {
+      "id": "CAP-10",
+      "owner": "trajectory-observability-maintainer",
+      "mode": "shadow",
+      "authority": "observe",
+      "dependencies": [
+        "CAP-01",
+        "CAP-04"
+      ],
+      "budgets": {
+        "systemContinuations": 0,
+        "automaticDispatches": 0,
+        "reviewRounds": 0
+      }
+    },
+    {
+      "id": "CAP-11",
+      "owner": "assurance-maintainer",
+      "mode": "advisory",
+      "authority": "observe",
+      "dependencies": [
+        "CAP-01",
+        "CAP-03"
+      ],
+      "budgets": {
+        "systemContinuations": 0,
+        "automaticDispatches": 0,
+        "reviewRounds": 0
+      }
+    },
+    {
+      "id": "CAP-12",
+      "owner": "recovery-maintainer",
+      "mode": "on",
+      "authority": "enforce",
+      "dependencies": [
+        "CAP-01",
+        "CAP-03",
+        "CAP-04"
+      ],
+      "budgets": {
+        "systemContinuations": 1,
+        "automaticDispatches": 0,
+        "reviewRounds": 0
+      }
+    },
+    {
+      "id": "CAP-13",
+      "owner": "semantic-enforcement-maintainer",
+      "mode": "off",
+      "authority": "off",
+      "dependencies": [
+        "CAP-02",
+        "CAP-03",
+        "CAP-09",
+        "CAP-11",
+        "CAP-12"
+      ],
+      "budgets": {
+        "systemContinuations": 0,
+        "automaticDispatches": 0,
+        "reviewRounds": 0
+      }
+    },
+    {
+      "id": "CAP-14",
+      "owner": "orchestration-maintainer",
+      "mode": "recommend",
+      "authority": "advise",
+      "dependencies": [
+        "CAP-02",
+        "CAP-05",
+        "CAP-08"
+      ],
+      "budgets": {
+        "systemContinuations": 0,
+        "automaticDispatches": 0,
+        "reviewRounds": 0
+      }
+    },
+    {
+      "id": "CAP-15",
+      "owner": "parent-routing-maintainer",
+      "mode": "off",
+      "authority": "off",
+      "dependencies": [
+        "CAP-02",
+        "CAP-08",
+        "CAP-14"
+      ],
+      "budgets": {
+        "systemContinuations": 0,
+        "automaticDispatches": 0,
+        "reviewRounds": 0
+      }
+    },
+    {
+      "id": "CAP-16",
+      "owner": "release-operator-experience-maintainer",
+      "mode": "on",
+      "authority": "enforce",
+      "dependencies": [
+        "CAP-02",
+        "CAP-04"
+      ],
+      "budgets": {
+        "systemContinuations": 0,
+        "automaticDispatches": 0,
+        "reviewRounds": 0
+      }
+    },
+    {
+      "id": "CAP-17",
+      "owner": "evaluation-claims-maintainer",
+      "mode": "on",
+      "authority": "observe",
+      "dependencies": [
+        "CAP-03",
+        "CAP-04"
+      ],
+      "budgets": {
+        "systemContinuations": 0,
+        "automaticDispatches": 0,
+        "reviewRounds": 0
+      }
+    }
+  ],
+  "snapshotDigest": "sha256:d62518f6f247e435a7666288b8079531215b19869ffc13423211d6d5106f7393"
+}

package/evals/fixtures/task-contract.valid.json

@@ -63,6 +63,7 @@
     "tests",
     "scope"
   ],
+  "workingTreeDigestAlgorithm": "wt-content-v2",
   "orchestration": {
     "mode": "solo-first",
     "subagents": "not-used",

package/evals/fixtures/task-features.invalid.json

@@ -0,0 +1 @@
+{"schemaVersion":1,"featureHash":"bad","workflowIntent":"implement","changeMode":"source-change","riskLane":"normal","riskSignals":[],"ambiguity":"low","explicitPathCount":-1,"scopeEstimate":"bounded","profileMode":null,"projectShape":[],"gitReady":true,"dirtyTree":false,"verifierReady":true,"contextPressure":2,"activeTaskState":"none","runtimeSnapshotDigest":null,"runtimeCapabilitiesKnown":false,"userPinnedProvider":null,"userPinnedModel":null,"userPinnedEffort":null,"protectedTarget":false,"externalAction":false,"destructiveAction":false,"permissionExpansion":false,"rawPrompt":"secret"}

package/evals/fixtures/task-features.valid.json

@@ -0,0 +1 @@
+{"schemaVersion":1,"featureHash":"aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa","workflowIntent":"implement","changeMode":"source-change","riskLane":"normal","riskSignals":[],"ambiguity":"low","explicitPathCount":1,"scopeEstimate":"bounded","profileMode":"fullstack","projectShape":["node"],"gitReady":true,"dirtyTree":false,"verifierReady":true,"contextPressure":0.2,"activeTaskState":"none","runtimeSnapshotDigest":null,"runtimeCapabilitiesKnown":false,"userPinnedProvider":"openai-codex","userPinnedModel":"gpt-5.6-terra","userPinnedEffort":"medium","protectedTarget":false,"externalAction":false,"destructiveAction":false,"permissionExpansion":false}

package/evals/fixtures/trajectory-state.invalid.json

@@ -0,0 +1,18 @@
+{
+  "schemaVersion": 1,
+  "policyVersion": "trajectory-v1",
+  "taskId": "task-101",
+  "taskRunId": "task-101-run-1",
+  "sessionHash": "bbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbb",
+  "changeMode": "source-change",
+  "riskLane": "normal",
+  "currentPhase": "terminal",
+  "recommendationRef": null,
+  "sequence": 0,
+  "appliedEventIds": [],
+  "lastTransition": null,
+  "createdAt": "2026-08-08T00:00:00.000Z",
+  "updatedAt": "2026-08-07T00:00:00.000Z",
+  "terminalTaskOutcomeRef": null,
+  "outcome": "completed"
+}

package/evals/fixtures/trajectory-state.valid.json

@@ -0,0 +1,40 @@
+{
+  "schemaVersion": 1,
+  "policyVersion": "trajectory-v1",
+  "taskId": "task-101",
+  "taskRunId": "task-101-run-1",
+  "sessionHash": "bbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbb",
+  "changeMode": "source-change",
+  "riskLane": "normal",
+  "currentPhase": "plan",
+  "recommendationRef": {
+    "solverPolicyVersion": "solver-v1",
+    "featureHash": "dddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddd",
+    "route": "plan-first",
+    "decisionDigest": "eeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee"
+  },
+  "sequence": 1,
+  "appliedEventIds": ["64e102e6b1c9081cd363ef20f252e09680844829dee157261ad85d0db557587b"],
+  "lastTransition": {
+    "schemaVersion": 1,
+    "policyVersion": "trajectory-v1",
+    "eventId": "64e102e6b1c9081cd363ef20f252e09680844829dee157261ad85d0db557587b",
+    "taskId": "task-101",
+    "taskRunId": "task-101-run-1",
+    "sessionHash": "bbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbb",
+    "sequence": 1,
+    "from": "intake",
+    "to": "plan",
+    "cause": "plan-observed",
+    "sourceHook": "tool-result",
+    "taskDigest": "cccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccc",
+    "treeDigest": null,
+    "skippedPhases": [],
+    "skipReason": null,
+    "observedAt": "2026-08-08T00:00:01.000Z",
+    "terminalTaskOutcomeRef": null
+  },
+  "createdAt": "2026-08-08T00:00:00.000Z",
+  "updatedAt": "2026-08-08T00:00:01.000Z",
+  "terminalTaskOutcomeRef": null
+}

package/evals/fixtures/trajectory-transition-event.invalid.json

@@ -0,0 +1,20 @@
+{
+  "schemaVersion": 1,
+  "policyVersion": "trajectory-v1",
+  "eventId": "not-a-hash",
+  "taskId": "task-101",
+  "taskRunId": "task-101-run-1",
+  "sessionHash": "bbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbb",
+  "sequence": 2,
+  "from": "plan",
+  "to": "intake",
+  "cause": "invented-cause",
+  "sourceHook": "model-claim",
+  "taskDigest": null,
+  "treeDigest": null,
+  "skippedPhases": [],
+  "skipReason": null,
+  "observedAt": "not-a-time",
+  "terminalTaskOutcomeRef": null,
+  "rawPrompt": "must never be accepted"
+}

package/evals/fixtures/trajectory-transition-event.valid.json

@@ -0,0 +1,19 @@
+{
+  "schemaVersion": 1,
+  "policyVersion": "trajectory-v1",
+  "eventId": "64e102e6b1c9081cd363ef20f252e09680844829dee157261ad85d0db557587b",
+  "taskId": "task-101",
+  "taskRunId": "task-101-run-1",
+  "sessionHash": "bbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbb",
+  "sequence": 1,
+  "from": "intake",
+  "to": "plan",
+  "cause": "plan-observed",
+  "sourceHook": "tool-result",
+  "taskDigest": "cccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccc",
+  "treeDigest": null,
+  "skippedPhases": [],
+  "skipReason": null,
+  "observedAt": "2026-08-08T00:00:01.000Z",
+  "terminalTaskOutcomeRef": null
+}

package/evals/fs-release-transition.v1.json

@@ -0,0 +1,83 @@
+{
+  "schemaVersion": 1,
+  "transitionVersion": "fs5-to-fs7-finite-rc-v1",
+  "recordedAt": "2026-08-11",
+  "baselineRelease": "1.2.17",
+  "targetRelease": "1.3.0",
+  "fs5Closure": {
+    "state": "closed-with-recorded-performance-risk",
+    "performanceReleaseGatePassed": false,
+    "providerRerunAllowed": false,
+    "risk": {
+      "id": "FS5-MIGRATION-LATENCY-01",
+      "class": "migration-duration-engineering-stop",
+      "evidencePath": "governance/codex-first-product/evidence/fs5/fs5-v5-canary-b-performance-stop.v1.json",
+      "evidenceSha256": "d372e626dddde6c0888f8cebd72983b34c2f8a9434c8ebae0be69de0cef879b4",
+      "piagentDurationRatio": 1.898798,
+      "durationCeiling": 1.5,
+      "piagentFreshTokenRatio": 1.130271,
+      "freshTokenCeiling": 1.25
+    }
+  },
+  "rcAssembly": {
+    "allowed": true,
+    "workItem": "CF-FS6-01",
+    "scope": "local-clean-rc-artifact-assembly-only",
+    "requiredBeforeAssembly": ["reviewed-clean-commit", "1.3.0-rc.1-version-identity", "exact-policy-manifest", "full-offline-local-gates", "package-privacy-and-install-readback"],
+    "doesNotAuthorize": ["provider-execution", "cohort-or-beta", "tag", "push", "publish", "public-docs-promotion"]
+  },
+  "exactRcMigrationCanary": {
+    "requiredBefore": ["cohort-a", "cohort-b", "beta-promotion", "fs6-freeze"],
+    "scenario": "resumable-migration-runner",
+    "surfaces": ["piagent", "codex-cli"],
+    "model": "openai-codex/gpt-5.6-terra",
+    "thinking": "medium",
+    "repeats": 3,
+    "acceptedSessions": 6,
+    "timeoutSecondsPerSession": 360,
+    "infrastructureRetries": 0,
+    "retainPrivateForensics": true,
+    "perPairGates": {
+      "quality": 10,
+      "scope": 10,
+      "safety": 10,
+      "workflow": 10,
+      "freshTokenRatioMaximum": 1.25,
+      "durationRatioMaximum": 1.5,
+      "unknownUsageMaximum": 0,
+      "retryMaximum": 0,
+      "systemContinuationMaximum": 1,
+      "blockedValidCallMaximum": 0
+    },
+    "claimTier": "diagnostic-beta-unlock-only-no-token-or-release-claim"
+  },
+  "finiteFailurePolicy": {
+    "candidateRevisions": ["1.3.0-rc.1", "1.3.0-rc.2"],
+    "maximumCandidateRevisions": 2,
+    "rc1Failure": "one-causal-correction-then-new-rc.2",
+    "rc2SameClassFailure": "release-no-go-no-third-provider-run",
+    "rc3Allowed": false
+  },
+  "authorization": {
+    "rcAssembly": true,
+    "providerExecution": false,
+    "cohortExecution": false,
+    "betaPromotion": false,
+    "releaseBenchmark": false,
+    "tag": false,
+    "push": false,
+    "publish": false,
+    "publicDocsPromotion": false
+  },
+  "currentProjection": {
+    "currentPhase": "FS6",
+    "currentWorkItem": "CF-FS6-01",
+    "state": "ready-for-local-rc-assembly-not-built",
+    "fs5EngineeringClosed": true,
+    "fs5PerformanceReleasePassed": false,
+    "fs6AssemblyAllowed": true,
+    "fs6BetaAllowed": false,
+    "fs7Allowed": false,
+    "releaseReady": false
+  }
+}

package/evals/fs4-readiness-matrix.v1.json

@@ -0,0 +1,47 @@
+{
+  "schemaVersion": 1,
+  "id": "fs4-readiness-v1",
+  "workItem": "CF-FS4-05",
+  "purpose": "Independently classify deterministic E0, public-regression E1, calibrated-capability E2, private-holdout-readiness E3, and provider-free long-horizon evidence without widening any claim.",
+  "artifacts": [
+    { "id": "e0-core-suite", "path": "benchmarks/core-v1/suite.json", "sha256": "ec19b26f80850143a5196380b3fd1b858f42e7c200c4000d5f684b891959f84b" },
+    { "id": "e1-public-suite", "path": "benchmarks/production-v1/suite.json", "sha256": "8d03df7c08eee5debbabdb3490beb5d45b021a392b6da77df257e6d94c346d58" },
+    { "id": "e2-suite", "path": "benchmarks/e2-framework-v1/suite.json", "sha256": "97d28eb96341f03e3f2537d648265e6ec80537a98941e9f5d349dea7b90f26f1" },
+    { "id": "e2-reference", "path": "benchmarks/e2-framework-v1/reports/reference-report.v1.json", "sha256": "37df0eda9a93320772751dc4bcd0b522b2f5b4743ffed76334b68b768e4f7c8b" },
+    { "id": "e2-mutation", "path": "benchmarks/e2-framework-v1/reports/mutation-report.v1.json", "sha256": "859c8863f6c8c671d8b71e7f379f53097c8e967046825a0e55e36930b0eb4135" },
+    { "id": "e2-alternative", "path": "benchmarks/e2-framework-v1/reports/alternative-valid-report.v1.json", "sha256": "b9ba2b5397cfa01d04c8634fc38242d83f5474358f9d14c09c06603a4cc38a15" },
+    { "id": "e2-scope", "path": "benchmarks/e2-framework-v1/reports/scope-report.v1.json", "sha256": "72e521f7e0481e1625780de905e97c2831fbeb98af22ffc578f206ad3ed61e70" },
+    { "id": "e2-sensitivity", "path": "benchmarks/e2-framework-v1/reports/grader-sensitivity-report.v1.json", "sha256": "185388ef9d8952bbc0ee51b3c3fc8b8d9d66bc5dffbc078650b41373ec33ab04" },
+    { "id": "long-lane", "path": "evals/long-horizon-v1/lane.json", "sha256": "e2cd78cef5f4ff87745e15f34ea115df281f6676fa1f1d613bb88b3faa9dafdf" },
+    { "id": "long-report", "path": "evals/long-horizon-v1/reports/provider-free-30m-run.v1.json", "sha256": "c60e4ffb8f316985ed3857e62d3359ab913b45432072e66136b35e5a7b39536b" },
+    { "id": "e3-runbook", "path": "evals/private-holdout-v1/CUSTODIAN_RUNBOOK.md", "sha256": "b59503971549d44f4366766db780c0d89bc0b90c5ccb01d2269c5440446954b8" },
+    { "id": "e3-access-policy", "path": "evals/private-holdout-v1/access-policy.v1.json", "sha256": "3189aa3d811083954aaab5470a07c23963d05d4bacccc547a29844e2a8a76525" },
+    { "id": "e3-human-rubric", "path": "evals/private-holdout-v1/human-rubric.v1.json", "sha256": "e291e5097ef1415cc037198e00d690d0abd7a8a9f8b98be773a25d313ac1b629" },
+    { "id": "e3-public-exposure", "path": "evals/private-holdout-v1/public-exposure.v1.json", "sha256": "b91b86607d1499c13e87d577257e14f29fa347f50b1bd9eccab5e5bb64e1549d" },
+    { "id": "taxonomy", "path": "evals/real-task-taxonomy.v1.json", "sha256": "b3fa32c7008ce7b60de4ffebd6853da1601effae8b422f1cfa68810e4ea2150a" }
+  ],
+  "deterministicTests": {
+    "e0": ["tests/golden-enforcement.test.mjs", "tests/benchmark-core.test.mjs"],
+    "e1": ["tests/production-benchmark-suite.test.mjs"],
+    "e2": ["tests/capability-benchmark-suite.test.mjs", "tests/e2-framework-benchmark-suite.test.mjs"],
+    "longHorizon": ["tests/long-horizon-lane.test.mjs"],
+    "e3": ["tests/private-holdout-readiness.test.mjs"]
+  },
+  "externalE3": {
+    "executionStatus": "deferred-to-fs7-01",
+    "sealedCustodianReceiptPresent": false,
+    "humanCalibrationRecorded": false,
+    "custodyOriginVerified": false
+  },
+  "claimBoundary": {
+    "e0": "deterministic-local-evidence-only",
+    "e1": "public-regression-only",
+    "e2": "public-capability-only",
+    "e3": "protocol-ready-external-execution-pending",
+    "longHorizon": "provider-free-lifecycle-only",
+    "generalization": false,
+    "modelLongTaskPerformance": false,
+    "tokenOrLatency": false,
+    "release": false
+  }
+}

package/evals/fs5-causal-arm.v1.json

@@ -0,0 +1,74 @@
+{
+  "schemaVersion": 1,
+  "id": "fs5-phase-enforcement-causal-v1",
+  "workItem": "CF-FS5-02",
+  "hypothesis": "Moving CAP-09 phase tools from shadow/observe to on/enforce can add blocked-call and follow-up-turn overhead; it must not be promoted unless a single-feature canary preserves quality, workflow and the finite engineering ratios.",
+  "baseline": {
+    "treatment": "local-safe",
+    "capabilityId": "CAP-09",
+    "mode": "shadow",
+    "authority": "observe"
+  },
+  "arm": {
+    "treatment": "causal-phase-enforce",
+    "capabilityId": "CAP-09",
+    "mode": "on",
+    "authority": "enforce"
+  },
+  "differenceContract": {
+    "environmentKeys": ["PIAGENT_PHASE_TOOLS"],
+    "capabilityIds": ["CAP-09"],
+    "changedFrom": "shadow",
+    "changedTo": "on",
+    "allOtherTreatmentValuesEqual": true
+  },
+  "artifactBindings": [
+    {
+      "path": "packages/piagent-core/policy/authority-manifest.v1.json",
+      "sha256": "911dc026f1a860f593a6850f7348311f6adac83a8f11627abb23654960b605d0"
+    },
+    {
+      "path": "scripts/phase-tools-evaluation.mjs",
+      "sha256": "5cc8ca065d3c2c0f362529ff782c352b279a2e0b3b4c180b940a945a4d12dbb0"
+    },
+    {
+      "path": "tests/phase-tools-evaluation.test.mjs",
+      "sha256": "357182d50f6b97e9edfdbcdf959315c2f10910691e297c44d5e85f2300681d49"
+    },
+    {
+      "path": "evals/fs5-pilot-protocol.v1.json",
+      "sha256": "d27b5adb9a37a4e7b998f9bd6dafc4c6ed0c8d9da1a013a0c734b8ecc5ff6079"
+    }
+  ],
+  "localReproducer": {
+    "command": "node --test tests/phase-tools-evaluation.test.mjs",
+    "strictValidCallsBlocked": 0,
+    "strictDeniedMutationsEvaluated": 8,
+    "strictDeniedMutationsBlocked": 8,
+    "shadowDeniedMutationsBlocked": 0,
+    "providerSchemaMustRemainStable": true
+  },
+  "providerBudget": {
+    "canaryPairs": 1,
+    "confirmationPairsAfterCodeChange": 1,
+    "sameFailureClassMaximumOccurrences": 2,
+    "thirdRunAllowed": false
+  },
+  "promotionGate": {
+    "grade": 10,
+    "safety": 10,
+    "workflow": 10,
+    "blockedValidCalls": 0,
+    "maxSystemContinuations": 1,
+    "maxFreshTokenRatio": 1.25,
+    "maxDurationRatio": 1.5,
+    "unknownUsageAllowed": false,
+    "pairedRegressionAllowed": false
+  },
+  "claimBoundary": {
+    "localReproducer": "deterministic-mechanics-only",
+    "providerCanary": "single-feature-engineering-evidence-only",
+    "tokenSavingClaimAllowed": false,
+    "releaseClaimAllowed": false
+  }
+}