@piagent/platform 1.5.5 → 1.6.1

package/benchmarks/deep-logic-v1/prompts/temporal-usage-billing-close.md

@@ -0,0 +1,123 @@
+Implement the temporal usage-billing close across:
+
+- `packages/billing/src/plan-timeline.js`
+- `packages/billing/src/close-period.js`
+- `apps/admin/src/billing-summary.js`
+
+The implementation must use only Node's standard library and must never mutate
+caller input. All accepted values are plain JSON values: reject sparse arrays,
+non-finite/unsafe numbers, non-plain object prototypes, accessors, symbol keys,
+unknown fields, and the keys/names `__proto__`, `prototype`, and `constructor`.
+Every rejection must throw `TypeError` before returning any partial result.
+
+## Period and plan timeline
+
+`period` contains exactly `{ start, end }`. Both values are canonical ISO-8601
+UTC strings (`new Date(value).toISOString() === value`) and define a non-empty
+half-open interval `[start, end)`.
+
+`normalizePlanTimeline(plans, period)` validates and returns a deep-cloned array
+sorted by `tenantId`, `effectiveAt`, then `id`. A plan contains exactly:
+
+```js
+{
+  id, tenantId, currency, effectiveAt,
+  meters: {
+    api: [
+      { upTo: 100, unitPriceMicros: 250000 },
+      { upTo: null, unitPriceMicros: 500000 }
+    ]
+  }
+}
+```
+
+- `id`, `tenantId`, and meter names are non-empty trimmed strings and cannot be
+  poison names. Plan ids are globally unique. Effective times are unique per
+  tenant. A tenant uses one three-letter uppercase currency for its whole
+  timeline.
+- `effectiveAt` is canonical ISO UTC. A plan becomes active at that exact
+  instant and remains active until the next plan for that tenant.
+- `meters` is a non-empty plain object. Its keys are returned in lexical order.
+  Every tier array is dense and non-empty. Each tier contains exactly `upTo`
+  and `unitPriceMicros`.
+- `upTo` values are cumulative positive safe-integer boundaries, strictly
+  increasing. Exactly the final tier has `upTo: null`. `unitPriceMicros` is a
+  non-negative safe integer. One unit price is expressed in one millionth of a
+  currency minor unit.
+
+## Usage close
+
+`closeUsagePeriod(input)` accepts exactly `{ period, plans, events }`. Events
+have globally unique, non-empty trimmed ids and one of these exact shapes:
+
+```js
+{ id, kind: "usage", tenantId, meter, at, units }
+{ id, kind: "reversal", tenantId, at, targetId }
+```
+
+Every event time must be canonical and inside the period. `units` is a positive
+safe integer. A reversal must reference an existing usage event for the same
+tenant at or before the reversal time. One usage can be reversed at most once;
+a reversal cannot target another reversal.
+
+Resolve all valid reversals before rating, independent of input order. Rate the
+remaining usage in ascending `at`, then `id` order. Select the latest plan whose
+`effectiveAt <= usage.at`. A missing active plan or a meter absent from that plan
+is invalid input: throw `TypeError`; never skip or silently omit that usage. Tier
+consumption is cumulative per `(tenantId, planId, meter)`, so it is isolated
+between tenants and meters and resets on a plan change. One usage event may span
+multiple tiers.
+
+Use `BigInt` internally for every multiplication and sum. Do not convert a
+charge through `Number`. Round an invoice from micro-minor units to minor units
+exactly once, at close, using round-half-to-even with divisor `1_000_000`.
+
+Return exactly:
+
+```js
+{
+  period: { start, end },
+  invoices: [{
+    tenantId,
+    currency,
+    lines: [{
+      eventId, planId, meter, at, units,
+      allocations: [{
+        tierIndex, units,
+        unitPriceMicros, // decimal string
+        chargeMicros     // decimal string
+      }],
+      chargeMicros       // decimal string
+    }],
+    subtotalMicros,      // decimal string
+    totalMinor,          // decimal string
+    digest               // lowercase SHA-256
+  }],
+  reversedEventIds
+}
+```
+
+Invoices are ordered by `tenantId`; lines use rating order; allocation order is
+tier order; reversed ids are lexical. Create an invoice only for a tenant with
+at least one non-reversed usage. `digest` is SHA-256 of canonical JSON for the
+invoice without `digest`: object keys recursively sorted and array order kept.
+
+`billingSummary(result)` returns a deterministic Terminal/WebUI summary. The
+lines are joined with the actual LF character (`"\n"`), never the two literal
+characters backslash and `n`. The first line is:
+
+```text
+period=<start>..<end>; invoices=<count>; reversed=<count>
+```
+
+Then emit one line per invoice, in result order:
+
+```text
+tenant=<JSON string>; currency=<currency>; lines=<count>; subtotalMicros=<value>; totalMinor=<value>; digest=<digest>
+```
+
+Reject malformed result objects rather than displaying a misleading summary.
+Add focused integration tests if useful; change only the declared source/test
+scope.
+
+Run `npm test` when complete; it includes a fixed public happy-path contract for this scenario.

package/benchmarks/deep-logic-v1/prompts/transactional-config-merge.md

@@ -11,3 +11,5 @@ Requirements:
 - The operation is transactional: any validation failure throws and yields no partial object.
 
 Use Node's standard library only. Add focused tests if useful.
+
+Run `npm test` when complete; it includes a fixed public happy-path contract for this scenario.

package/benchmarks/deep-logic-v1/suite.json

@@ -1,11 +1,45 @@
 {
-  "schemaVersion": 1,
+  "schemaVersion": 2,
   "id": "deep-logic-v1",
   "title": "Piagent Deep Logic V1",
-  "description": "Paired current-model tasks with interacting invariants across event reconciliation, scheduling, layered policy, graph context, stream recovery, and transactional configuration.",
+  "description": "Paired current-model tasks with interacting invariants across event reconciliation, scheduling, layered policy, graph context, stream recovery, transactional configuration, and temporal usage billing.",
   "profile": "node-typescript",
-  "defaultRepeats": 2,
+  "defaultRepeats": 3,
   "timeoutSeconds": 1200,
+  "executionContract": {
+    "surfaces": ["piagent", "codex-cli"],
+    "model": "openai-codex/gpt-5.6-luna",
+    "thinking": "medium",
+    "codexMode": "controlled"
+  },
+  "assurance": {
+    "taskSource": "maintainer-authored-deep-work-archetypes",
+    "visibility": "public-suite-with-run-private-oracles",
+    "claimTier": "capability",
+    "generatedVariants": true,
+    "familyDisjointSplit": false,
+    "reviewed": true,
+    "refreshedAt": "2026-08-21"
+  },
+  "releaseGate": {
+    "minimumQualityScore": 9.5,
+    "minimumSafetyScore": 10,
+    "minimumReliabilityScore": 9.5,
+    "minimumWorkflowScore": 9.5,
+    "minimumCategoryScore": 9.5,
+    "minimumOutcomeScoreExclusive": 9.5,
+    "minimumPairedScenarios": 7,
+    "minimumComparableEfficiencyScenarios": 7,
+    "minimumRepeats": 3,
+    "maximumInfrastructureRetries": 0,
+    "maximumFreshTokenRatioUpper95": 0.8,
+    "maximumDurationRatioUpper95": 1.1,
+    "primaryEfficiencyEstimand": "failure-aware-family-ratio",
+    "requireEfficiencyClaim": true,
+    "requireFullSuiteForClaim": true,
+    "requireStableProviderWireSurface": true,
+    "requireCausalContextReceipt": true
+  },
   "scenarios": [
     {
       "id": "revision-event-reconciliation",
@@ -90,6 +124,20 @@
       "grader": "grade.mjs",
       "variantGenerator": "variant.mjs",
       "allowedChanges": ["src/config-transaction.js", "test/**"]
+    },
+    {
+      "id": "temporal-usage-billing-close",
+      "title": "Temporal usage billing close with exact audit",
+      "kind": "source-change",
+      "category": "financial-correctness",
+      "difficulty": "large",
+      "profile": "fullstack",
+      "lifecycle": "cold-start",
+      "fixture": "project",
+      "prompt": "prompts/temporal-usage-billing-close.md",
+      "grader": "grade.mjs",
+      "variantGenerator": "variant.mjs",
+      "allowedChanges": ["packages/billing/src/**", "apps/admin/src/billing-summary.js", "test/**"]
     }
   ]
 }

package/benchmarks/deep-logic-v1/variant.mjs

@@ -20,8 +20,29 @@ switch (scenarioId) {
   }; break;
   case "layered-policy-resolution": oracle.graderData = { protectedPath: `private/${token("vault")}/secret.json`, operation: token("read") }; break;
   case "budgeted-context-graph": oracle.graderData = { term: token("resolver"), budget: integer(8, 14) }; break;
-  case "resumable-stream-assembly": oracle.graderData = { first: token("hello"), second: token("world"), message: token("message") }; break;
+  case "resumable-stream-assembly": {
+    const astral = ["🙂", "🧠", "🚀"][integer(0, 2)];
+    oracle.graderData = {
+      first: token("hello"),
+      second: token("world"),
+      third: token("chunk"),
+      message: token("message"),
+      other: token("message"),
+      start: integer(2, 24),
+      astral
+    };
+    break;
+  }
   case "transactional-config-merge": oracle.graderData = { key: token("feature"), value: token("enabled") }; break;
+  case "temporal-usage-billing-close": oracle.graderData = {
+    tenantA: token("tenant"),
+    tenantB: token("tenant"),
+    meter: token("meter"),
+    plans: [token("plan"), token("plan"), token("plan")],
+    firstLimit: integer(4, 12),
+    secondLimit: integer(15, 28),
+    prices: [integer(110_000, 390_000), integer(410_000, 790_000), integer(810_000, 1_250_000)]
+  }; break;
   default: throw new Error(`unsupported deep logic scenario ${scenarioId}`);
 }
 

package/benchmarks/production-v1/spend-control.v1.json

@@ -0,0 +1,92 @@
+{
+  "schemaVersion": 1,
+  "suiteId": "production-v1",
+  "purpose": "Provider-spend control before the one claim-eligible 108-session token-quality-continuity result.",
+  "rootSeed": "production-v1-40-gate-v1-1370",
+  "execution": {
+    "surfaces": ["piagent", "codex-cli"],
+    "model": "openai-codex/gpt-5.6-luna",
+    "thinking": "medium",
+    "repeats": 3,
+    "infrastructureRetries": 0,
+    "stopAfterFailedPair": true
+  },
+  "stages": [
+    { "id": "S0", "cumulativeSessions": 0, "newSessions": 0, "mode": "preflight-only", "claimEligible": false },
+    { "id": "S12", "cumulativeSessions": 12, "newSessions": 12, "mode": "execute-and-pause", "claimEligible": false },
+    { "id": "S36", "cumulativeSessions": 36, "newSessions": 24, "mode": "resume-and-pause", "claimEligible": false },
+    { "id": "S72", "cumulativeSessions": 72, "newSessions": 36, "mode": "resume-and-pause", "claimEligible": false },
+    { "id": "S108", "cumulativeSessions": 108, "newSessions": 36, "mode": "resume-and-finalize", "claimEligible": true }
+  ],
+  "earlyDetection": {
+    "expectedFirstRepeatScenarioIds": [
+      "cli-double-dash",
+      "destructive-history-refusal",
+      "tenant-role-authorization",
+      "expiry-boundary",
+      "quoted-csv",
+      "unicode-search"
+    ],
+    "requiredCategoryCoverage": ["backend", "frontend", "data", "platform", "reliability", "security"],
+    "requiredLifecycleCoverage": ["steady-state", "cold-start"],
+    "requiredDifficultyCoverage": ["small", "medium", "large"]
+  },
+  "productionGuards": {
+    "subagents": {
+      "required": true,
+      "maximumSessionsPerAttempt": 1,
+      "maximumTrafficShare": 0.05,
+      "requireExactAllAttemptEvidence": true,
+      "requireExplainedSessionCount": true
+    },
+    "partialFreshSpend": {
+      "requiredFromCumulativeSessions": 12,
+      "maximumPooledFreshRatio": 1.1,
+      "maximumObservedFamilyFreshRatio": 1.25,
+      "includeExactFailedAttempts": true
+    },
+    "providerFreeEvidence": {
+      "requiredBeforeFirstPaidSession": true,
+      "requiredLaneIds": ["runtime-conformance-v1", "long-horizon-v1", "webui-parity-v1"],
+      "requireCleanCommitAndTreeBinding": true,
+      "requireProductionConfigurationBinding": true,
+      "requireRunnerAndLaneConfigurationDigests": true
+    }
+  },
+  "enforcement": {
+    "runnerConsumesThisArtifact": true,
+    "runnerBindsFrozenSeedAndExecution": true,
+    "runnerWritesAtomicPrivateStageDiagnostic": true,
+    "runnerBlocksProductionResumeWhenStageDiagnosticFails": true,
+    "durableStageStateSurvivesMissingPauseMarker": true,
+    "completeLedgerFinalizesWithoutProviderPreflight": true,
+    "stageDiagnosticArtifact": "stage-diagnostic.json",
+    "automaticTerminalOrAbortRules": [
+      "dirty-release-source-before-auth-tool-or-provider-preflight",
+      "piagent-unresolved-or-outcome-score-at-or-below-floor-via-stop-after-failed-pair",
+      "infrastructure-error-or-unknown-provider-attempt-usage-via-runner-abort",
+      "candidate-provenance-drift-via-execution-guard"
+    ],
+    "stageAdvanceRules": [
+      "clean-release-source",
+      "no-baseline-pass-piagent-fail",
+      "no-observed-paired-grade-regression",
+      "all-observed-piagent-outcomes-remain-above-the-release-floor",
+      "no-model-thinking-or-provider-wire-parity-drift",
+      "complete-privacy-safe-causal-context-receipt-with-closed-session-lifecycle",
+      "source-bound-provider-free-runtime-long-horizon-and-webui-parity-receipts",
+      "exact-subagent-budget-at-most-one-session-per-attempt-and-five-percent-traffic",
+      "partial-stage-pooled-fresh-ratio-at-most-1.10-and-every-observed-family-at-most-1.25",
+      "no-accepted-usage-completeness-failure",
+      "no-infrastructure-retry",
+      "no-unknown-provider-attempt-usage"
+    ],
+    "nonBlockingDiagnostics": [
+      "normalized-api-equivalent-text-token-cost",
+      "duration",
+      "host-load"
+    ],
+    "operatorRule": "The commands pass this artifact's frozen values explicitly. S0 binds clean-source runtime-conformance, long-horizon, and WebUI-parity receipts before paid work. Every partial paid stage from S12 onward hard-stops when pooled fresh traffic exceeds 1.10, any observed family exceeds 1.25, or exact subagent evidence exceeds one child session per attempt or 5% of candidate traffic. Cost, duration, and host load remain observational. Only the complete S108 ledger may decide the 0.60 upper-95 fresh-token claim."
+  },
+  "claimBoundary": "Only S108 may be evaluated for the production-v1 0.60 upper-95 fresh-token public-regression claim. S12, S36 and S72 enforce catastrophic fresh-spend and subagent safety ceilings but cannot establish the claim. Duration, host load, and normalized cost are reported but cannot block or support this claim."
+}

package/benchmarks/production-v1/suite.json

@@ -6,6 +6,38 @@
   "profile": "node-typescript",
   "defaultRepeats": 3,
   "timeoutSeconds": 900,
+  "executionContract": {
+    "surfaces": ["piagent", "codex-cli"],
+    "model": "openai-codex/gpt-5.6-luna",
+    "thinking": "medium",
+    "codexMode": "controlled"
+  },
+  "pricingSnapshot": {
+    "schemaVersion": 1,
+    "id": "openai-gpt-5-6-luna-2026-08-22",
+    "model": "openai-codex/gpt-5.6-luna",
+    "currency": "USD",
+    "unitTokens": 1000000,
+    "rates": {
+      "freshInput": 0.2,
+      "cachedInput": 0.02,
+      "output": 1.2
+    },
+    "cacheWrite": {
+      "basis": "fresh-input",
+      "multiplier": 1.25
+    },
+    "longContext": {
+      "thresholdInputTokens": 272000,
+      "condition": "per-request-input-greater-than",
+      "inputMultiplier": 2,
+      "outputMultiplier": 1.5
+    },
+    "source": {
+      "url": "https://developers.openai.com/api/docs/models/gpt-5.6-luna",
+      "retrievedAt": "2026-08-22"
+    }
+  },
   "assurance": {
     "taskSource": "maintainer-authored-production-archetypes",
     "visibility": "public-suite-with-run-private-oracles",
@@ -23,10 +55,21 @@
     "minimumCategoryScore": 9.5,
     "minimumOutcomeScoreExclusive": 9.5,
     "minimumPairedScenarios": 18,
-    "minimumComparableEfficiencyScenarios": 12,
+    "minimumComparableEfficiencyScenarios": 18,
     "minimumRepeats": 3,
-    "maximumFreshTokenRatioUpper95": 1,
-    "requireEfficiencyClaim": true
+    "maximumInfrastructureRetries": 0,
+    "primaryEfficiencyEstimand": "fixed-workload-family-ratio",
+    "maximumFreshTokenRatioUpper95": 0.6,
+    "requireEfficiencyClaim": true,
+    "requireNormalizedCostClaim": false,
+    "requireHostReadinessForClaim": false,
+    "requireFullSuiteForClaim": true,
+    "requireStableProviderWireSurface": true,
+    "requireCausalContextReceipt": true,
+    "requireProviderFreeEvidence": true,
+    "requireSubagentBudget": true,
+    "maximumSubagentSessionsPerAttempt": 1,
+    "maximumSubagentTrafficShare": 0.05
   },
   "scenarios": [
     {

package/catalog/capabilities.json

@@ -26,8 +26,8 @@
           "path": "evals/scenarios/capability-resolution.json"
         },
         {
-          "byteSize": 6271,
-          "digest": "sha256:90dafc5b6c0bd555b932d89b10785b9eb50104ed22dcb06cfb8a2d306cc26d2f",
+          "byteSize": 7158,
+          "digest": "sha256:8edb1cce6716291c8ca642d3d2392c8f6c92e725ca502e93e06ab45abe541a79",
           "id": "base-policy",
           "kind": "policies",
           "path": "packages/piagent-core/policies/base-policy.json"
@@ -40,8 +40,8 @@
           "path": "packages/piagent-core/prompts/commit.md"
         },
         {
-          "byteSize": 1605,
-          "digest": "sha256:1c2ae0893f131ca0c519f0954845f2b4a8e61182360aac245134102dc3324643",
+          "byteSize": 1413,
+          "digest": "sha256:cb6e7bad1305c7b61d34489cbb5d066ca99855dcfc846dee1668d16d17be8a2b",
           "id": "plan",
           "kind": "prompts",
           "path": "packages/piagent-core/prompts/plan.md"
@@ -54,15 +54,15 @@
           "path": "packages/piagent-core/prompts/pr.md"
         },
         {
-          "byteSize": 1314,
-          "digest": "sha256:e1e9caf68957748bebf6ac2d6294905226b298b2734434db8341b35664794388",
+          "byteSize": 1375,
+          "digest": "sha256:bd474a18754c3ab835951d802158c5a19bd5e92f66cffd0dd6b560adcadb795d",
           "id": "review",
           "kind": "prompts",
           "path": "packages/piagent-core/prompts/review.md"
         },
         {
-          "byteSize": 2371,
-          "digest": "sha256:bfe8374f15e06d410eec7ab2c82f85ef011fc7d514a92f23720179b497910a44",
+          "byteSize": 2796,
+          "digest": "sha256:8f1552bac55dce1399f3e65cce0333205b2002cc37d4366d69e25b0b4d5f09f9",
           "id": "task",
           "kind": "prompts",
           "path": "packages/piagent-core/prompts/task.md"
@@ -82,29 +82,29 @@
           "path": "packages/piagent-core/skills/piagent-ops/SKILL.md"
         },
         {
-          "byteSize": 2024,
-          "digest": "sha256:9c6aa91663a6614b72a4fb8c6c4a07c07a466c0889fcfe45528ae3e3f62bac90",
+          "byteSize": 2223,
+          "digest": "sha256:18f865d816c5a5cee50bf3b2b407c23a2e67e1e92d929001a0f24c04d0c20e03",
           "id": "piagent-planner",
           "kind": "subagents",
           "path": "packages/piagent-core/subagents/piagent-planner.md"
         },
         {
-          "byteSize": 1951,
-          "digest": "sha256:6a02cbc7904ea6d8de64aa776ac21ccfbf3c9cc468c4c1c7650e08374ab9e949",
+          "byteSize": 2091,
+          "digest": "sha256:4e514d6480a618d610d4f520af93fa4e151ef60b42b30f5cf94da1a9c96d6ead",
           "id": "piagent-reviewer",
           "kind": "subagents",
           "path": "packages/piagent-core/subagents/piagent-reviewer.md"
         },
         {
-          "byteSize": 1575,
-          "digest": "sha256:ecd432c57c99783ab76844d0a35cccad377e4cd265674577742ded48ea3ffbb4",
+          "byteSize": 1714,
+          "digest": "sha256:37d9638bfbca8d50f86cbee3fee4db3ba9197bf06004d3a7bba08fc2f8249549",
           "id": "piagent-scout",
           "kind": "subagents",
           "path": "packages/piagent-core/subagents/piagent-scout.md"
         },
         {
-          "byteSize": 2146,
-          "digest": "sha256:e8e18e1ef3a43db2acd559c40a99b31d8cd994cbfaaf1adc5ef9d2cc5e3721a6",
+          "byteSize": 2210,
+          "digest": "sha256:aa18041470aae1a2b600cef1c09ddbfcd8b86e1efb6858c1f460e46292120e9d",
           "id": "piagent-worker",
           "kind": "subagents",
           "path": "packages/piagent-core/subagents/piagent-worker.md"
@@ -151,8 +151,8 @@
       },
       "artifacts": [
         {
-          "byteSize": 1950,
-          "digest": "sha256:84ac29dcbe0c692070be302707424ea16130a5938cff309c5d7755204b6c2dc1",
+          "byteSize": 2378,
+          "digest": "sha256:35d6757d0a8d2f8644b967fe82645fba3de430b53f23b5577c2e5e104dcccdc1",
           "id": "web-frontend",
           "kind": "adapters",
           "path": "adapters/web-frontend/profile.json"

package/docs/authority-manifest.md

@@ -170,7 +170,7 @@ directory sibling, unresolved test, denied carrier, failed call, or successful
 no-op locks the opportunity and invalidates reusable review credit. A real
 mutation must be followed by the task's exact verifier on the new tree before
 completion. Retryable infrastructure failure may retry that verifier once on
-the same digest, and a high-confidence in-scope failure may open one final
+the same digest, and a high-confidence evidence-backed failure may open one final
 mutation-bound correction inside the already consumed continuation. Every
 other failure, repeated signature, exhausted budget, stale/corrupt state, or
 unsupported semantic shape produces a deterministic handoff rather than a new