@piagent/platform 1.2.17 → 1.3.0-ie.9

package/benchmarks/e2-framework-v1/prompts/hono-accessible-search.md

@@ -0,0 +1,9 @@
+Implement the real Hono search page in `apps/web/src/search-app.js`. Keep changes inside the declared source and tests.
+
+- [S1] `normalizeQuery` must stringify non-nullish values, normalize NFD combining marks, trim, collapse Unicode whitespace, and lowercase.
+- [S2] `createSearchApp({ items })` returns a Hono app with `GET /search`; match normalized names and string tags in input order without mutation.
+- [S3] The optional `limit` query defaults to 20 and accepts safe integers 1 through 100; invalid values return 400 JSON rather than an exception page.
+- [S4] Return HTML with a labeled search form and `<ul aria-label="Search results">`; escape `& < > " '` in query, ids, and names.
+- [S5] Empty results still render the accessible empty list, and caller arrays/objects remain byte-for-value unchanged.
+
+Run `npm test` when complete.

package/benchmarks/e2-framework-v1/prompts/hono-tenant-api.md

@@ -0,0 +1,9 @@
+Implement the tenant user route in `apps/api/src/tenant-app.js` using the pinned Hono framework already vendored in this repository. Keep changes inside the declared source and tests.
+
+- [A1] `createTenantApp({ users })` returns a Hono app and registers `GET /tenants/:tenantId/users/:userId`.
+- [A2] Only `owner` or `admin` callers whose trimmed `x-tenant-id` exactly matches the trimmed path tenant may read the route; every other caller receives 403 without user data.
+- [A3] Validate `users` as an array and trim/reject empty tenant, user, header tenant, and role values with a 400 JSON response.
+- [A4] An active matching user returns 200 and exactly `{ id, tenantId, name }`; a missing or inactive user returns 404.
+- [A5] Do not mutate the caller's users array or user objects and do not return internal object references.
+
+Run `npm test` when complete.

package/benchmarks/e2-framework-v1/prompts/sqlite-resumable-inventory.md

@@ -0,0 +1,9 @@
+Implement `migrateInventory(db, options)` in `packages/migration/src/inventory.js` using the provided Node `DatabaseSync` instance. Keep changes inside the declared source and tests.
+
+- [M1] Accept only a `DatabaseSync`-compatible object. Legacy `inventory(id,name,quantity)` rows migrate to `inventory_v2(id,label,quantity)` and metadata version 2.
+- [M2] Trim ids/names, require non-empty values, and convert quantity to a safe non-negative integer. Any invalid row aborts the whole transaction.
+- [M3] The first successful run returns `{ version: 2, migrated: count }`; later runs are idempotent and return `{ version: 2, migrated: 0 }` without duplicating data.
+- [M4] `options.crashAfter` is an optional positive safe integer used to inject a failure after that many inserts. A crash leaves no partial v2 schema/data/version and a later normal call completes.
+- [M5] Preserve the legacy table and its rows; use prepared statements and a transaction rather than rebuilding from JavaScript-only state.
+
+Run `npm test` when complete.

package/benchmarks/e2-framework-v1/prompts/workspace-policy-rollout.md

@@ -0,0 +1,9 @@
+Implement the feature policy across the three declared workspace modules. Keep changes inside the declared source and tests.
+
+- [R1] `normalizePolicy` returns a new exact `{ enabled, percentage, tenants }` object, trims and first-seen-deduplicates tenants, and never mutates input.
+- [R2] Reject null/array/non-object policy, percentage outside safe integer 0..100, non-array tenants, non-string tenants, and empty trimmed tenants with `TypeError`.
+- [R3] `isEnabled` returns false for disabled or missing subjects, enables listed tenants, otherwise uses the exclusive `bucket < percentage` rule and validates bucket as safe integer 0..99.
+- [R4] `evaluateFeature` must import the policy package and return exact reasons `disabled`, `tenant-override`, `percentage`, or `not-eligible` while preserving validation errors.
+- [R5] `renderPolicySummary` must import and normalize through the policy package and return `enabled=<bool>; percentage=<n>; tenants=<comma-list>`.
+
+Run `npm test` when complete.

package/benchmarks/e2-framework-v1/references/hono-accessible-search/apps/web/src/search-app.js

@@ -0,0 +1,29 @@
+import { Hono } from "../../../vendor/hono/dist/index.js";
+
+export function normalizeQuery(value) {
+  return String(value ?? "").normalize("NFD").replace(/\p{M}+/gu, "").trim().replace(/\s+/gu, " ").toLowerCase();
+}
+
+function escapeHtml(value) {
+  return String(value).replaceAll("&", "&amp;").replaceAll("<", "&lt;").replaceAll(">", "&gt;").replaceAll('"', "&quot;").replaceAll("'", "&#39;");
+}
+
+export function createSearchApp({ items = [] } = {}) {
+  const app = new Hono();
+  app.get("/search", (context) => {
+    const rawQuery = context.req.query("q") ?? "";
+    const rawLimit = context.req.query("limit");
+    const limit = rawLimit === undefined ? 20 : Number(rawLimit);
+    if (!Number.isSafeInteger(limit) || limit < 1 || limit > 100 || !Array.isArray(items)) {
+      return context.json({ error: "invalid-request" }, 400);
+    }
+    const query = normalizeQuery(rawQuery);
+    const results = items.filter((item) => {
+      const fields = [item?.name, ...(Array.isArray(item?.tags) ? item.tags.filter((tag) => typeof tag === "string") : [])];
+      return fields.some((field) => normalizeQuery(field).includes(query));
+    }).slice(0, limit);
+    const rows = results.map((item) => `<li data-id="${escapeHtml(item.id)}">${escapeHtml(item.name)}</li>`).join("");
+    return context.html(`<form aria-label="Search"><label>Query <input name="q" value="${escapeHtml(rawQuery)}"></label></form><ul aria-label="Search results">${rows}</ul>`);
+  });
+  return app;
+}

package/benchmarks/e2-framework-v1/references/hono-tenant-api/apps/api/src/tenant-app.js

@@ -0,0 +1,22 @@
+import { Hono } from "../../../vendor/hono/dist/index.js";
+
+const text = (value) => typeof value === "string" ? value.trim() : "";
+
+export function createTenantApp({ users = [] } = {}) {
+  const app = new Hono();
+  app.get("/tenants/:tenantId/users/:userId", (context) => {
+    if (!Array.isArray(users)) return context.json({ error: "invalid-request" }, 400);
+    const tenantId = text(context.req.param("tenantId"));
+    const userId = text(context.req.param("userId"));
+    const callerTenant = text(context.req.header("x-tenant-id"));
+    const role = text(context.req.header("x-role"));
+    if (!tenantId || !userId || !callerTenant || !role) return context.json({ error: "invalid-request" }, 400);
+    if (callerTenant !== tenantId || !["owner", "admin"].includes(role)) {
+      return context.json({ error: "forbidden" }, 403);
+    }
+    const user = users.find((item) => item && text(item.id) === userId && text(item.tenantId) === tenantId);
+    if (!user || user.active !== true) return context.json({ error: "not-found" }, 404);
+    return context.json({ id: text(user.id), tenantId: text(user.tenantId), name: String(user.name) });
+  });
+  return app;
+}

package/benchmarks/e2-framework-v1/references/sqlite-resumable-inventory/packages/migration/src/inventory.js

@@ -0,0 +1,34 @@
+function compatible(db) {
+  return db && typeof db.exec === "function" && typeof db.prepare === "function";
+}
+
+export function migrateInventory(db, options = {}) {
+  if (!compatible(db)) throw new TypeError("db must be DatabaseSync-compatible");
+  if (!options || Array.isArray(options) || typeof options !== "object") throw new TypeError("options must be an object");
+  const crashAfter = options.crashAfter;
+  if (crashAfter !== undefined && (!Number.isSafeInteger(crashAfter) || crashAfter <= 0)) throw new TypeError("crashAfter must be a positive safe integer");
+  const metadataExists = db.prepare("SELECT 1 AS ok FROM sqlite_master WHERE type='table' AND name='migration_metadata'").get();
+  if (metadataExists && db.prepare("SELECT value FROM migration_metadata WHERE key='inventory-version'").get()?.value === 2) return { version: 2, migrated: 0 };
+  db.exec("BEGIN IMMEDIATE");
+  try {
+    db.exec("CREATE TABLE inventory_v2 (id TEXT PRIMARY KEY, label TEXT NOT NULL, quantity INTEGER NOT NULL); CREATE TABLE migration_metadata (key TEXT PRIMARY KEY, value INTEGER NOT NULL)");
+    const rows = db.prepare("SELECT id, name, quantity FROM inventory ORDER BY rowid").all();
+    const insert = db.prepare("INSERT INTO inventory_v2 (id, label, quantity) VALUES (?, ?, ?)");
+    let migrated = 0;
+    for (const row of rows) {
+      const id = typeof row.id === "string" ? row.id.trim() : "";
+      const label = typeof row.name === "string" ? row.name.trim() : "";
+      const quantity = Number(row.quantity);
+      if (!id || !label || !Number.isSafeInteger(quantity) || quantity < 0) throw new TypeError("invalid inventory row");
+      insert.run(id, label, quantity);
+      migrated += 1;
+      if (migrated === crashAfter) throw new Error("injected migration crash");
+    }
+    db.prepare("INSERT INTO migration_metadata (key, value) VALUES ('inventory-version', 2)").run();
+    db.exec("COMMIT");
+    return { version: 2, migrated };
+  } catch (error) {
+    db.exec("ROLLBACK");
+    throw error;
+  }
+}

package/benchmarks/e2-framework-v1/references/workspace-policy-rollout/apps/admin/src/summary.js

@@ -0,0 +1,6 @@
+import { normalizePolicy } from "../../../packages/policy/src/rollout.js";
+
+export function renderPolicySummary(policy) {
+  const value = normalizePolicy(policy);
+  return `enabled=${value.enabled}; percentage=${value.percentage}; tenants=${value.tenants.join(",")}`;
+}

package/benchmarks/e2-framework-v1/references/workspace-policy-rollout/packages/feature-api/src/evaluate.js

@@ -0,0 +1,10 @@
+import { isEnabled, normalizePolicy } from "../../policy/src/rollout.js";
+
+export function evaluateFeature(policy, subject) {
+  const normalized = normalizePolicy(policy);
+  if (!normalized.enabled) return { allowed: false, reason: "disabled" };
+  if (!subject) return { allowed: false, reason: "not-eligible" };
+  const allowed = isEnabled(normalized, subject);
+  if (normalized.tenants.includes(typeof subject.tenantId === "string" ? subject.tenantId.trim() : "")) return { allowed: true, reason: "tenant-override" };
+  return { allowed, reason: allowed ? "percentage" : "not-eligible" };
+}

package/benchmarks/e2-framework-v1/references/workspace-policy-rollout/packages/policy/src/rollout.js

@@ -0,0 +1,20 @@
+export function normalizePolicy(input) {
+  if (!input || Array.isArray(input) || typeof input !== "object") throw new TypeError("policy must be an object");
+  if (!Number.isSafeInteger(input.percentage) || input.percentage < 0 || input.percentage > 100) throw new TypeError("invalid percentage");
+  if (!Array.isArray(input.tenants)) throw new TypeError("tenants must be an array");
+  const tenants = [];
+  for (const tenant of input.tenants) {
+    if (typeof tenant !== "string" || !tenant.trim()) throw new TypeError("invalid tenant");
+    const value = tenant.trim();
+    if (!tenants.includes(value)) tenants.push(value);
+  }
+  return { enabled: Boolean(input.enabled), percentage: input.percentage, tenants };
+}
+
+export function isEnabled(policy, subject) {
+  const normalized = normalizePolicy(policy);
+  if (!normalized.enabled || !subject) return false;
+  if (!Number.isSafeInteger(subject.bucket) || subject.bucket < 0 || subject.bucket > 99) throw new TypeError("invalid bucket");
+  const tenantId = typeof subject.tenantId === "string" ? subject.tenantId.trim() : "";
+  return normalized.tenants.includes(tenantId) || subject.bucket < normalized.percentage;
+}

package/benchmarks/e2-framework-v1/reports/alternative-valid-report.v1.json

@@ -0,0 +1,10 @@
+{
+  "schemaVersion": 1,
+  "suiteId": "e2-framework-v1",
+  "kind": "alternative-valid",
+  "scenarioCount": 4,
+  "structurallyDistinct": true,
+  "expectedPassingScenarios": ["hono-tenant-api", "hono-accessible-search", "sqlite-resumable-inventory", "workspace-policy-rollout"],
+  "expectedScore": 10,
+  "providerUsed": false
+}

package/benchmarks/e2-framework-v1/reports/grader-sensitivity-report.v1.json

@@ -0,0 +1,15 @@
+{
+  "schemaVersion": 1,
+  "suiteId": "e2-framework-v1",
+  "kind": "grader-sensitivity",
+  "rubricCheckCount": 20,
+  "seededScenarioCount": 4,
+  "seededExpectedPassing": 0,
+  "referenceExpectedPassing": 4,
+  "alternativeExpectedPassing": 4,
+  "targetedMutationsExpectedKilled": 20,
+  "outsideScopeExpectedRejected": 4,
+  "claimTier": "capability",
+  "releaseClaimAllowed": false,
+  "providerUsed": false
+}

package/benchmarks/e2-framework-v1/reports/mutation-report.v1.json

@@ -0,0 +1,15 @@
+{
+  "schemaVersion": 1,
+  "suiteId": "e2-framework-v1",
+  "kind": "mutation",
+  "checkCount": 20,
+  "mutationCount": 20,
+  "expectedKilled": 20,
+  "providerUsed": false,
+  "checkIds": [
+    "tenant-hono-route", "tenant-authorization", "tenant-validation", "tenant-response-contract", "tenant-immutability",
+    "search-normalization", "search-hono-route", "search-limit-contract", "search-accessible-escaping", "search-empty-immutable",
+    "inventory-schema-migration", "inventory-validation-atomicity", "inventory-idempotency", "inventory-crash-resume", "inventory-preserves-legacy",
+    "policy-normalized-shape", "policy-validation", "policy-enablement", "policy-api-reasons", "policy-admin-summary"
+  ]
+}

package/benchmarks/e2-framework-v1/reports/reference-report.v1.json

@@ -0,0 +1,9 @@
+{
+  "schemaVersion": 1,
+  "suiteId": "e2-framework-v1",
+  "kind": "reference",
+  "scenarioCount": 4,
+  "expectedPassingScenarios": ["hono-tenant-api", "hono-accessible-search", "sqlite-resumable-inventory", "workspace-policy-rollout"],
+  "expectedScore": 10,
+  "providerUsed": false
+}

package/benchmarks/e2-framework-v1/reports/scope-report.v1.json

@@ -0,0 +1,11 @@
+{
+  "schemaVersion": 1,
+  "suiteId": "e2-framework-v1",
+  "kind": "scope",
+  "scenarioCount": 4,
+  "inScopeReferenceExpected": true,
+  "outsideScopeProbe": "README.md",
+  "outsideScopeExpected": false,
+  "vendorMutationAllowed": false,
+  "providerUsed": false
+}

package/benchmarks/e2-framework-v1/rubric.json

@@ -0,0 +1,33 @@
+{
+  "schemaVersion": 1,
+  "scenarios": {
+    "hono-tenant-api": [
+      { "id": "tenant-hono-route", "clause": "A1", "weight": 1, "critical": true },
+      { "id": "tenant-authorization", "clause": "A2", "weight": 1, "critical": true },
+      { "id": "tenant-validation", "clause": "A3", "weight": 1, "critical": true },
+      { "id": "tenant-response-contract", "clause": "A4", "weight": 1, "critical": true },
+      { "id": "tenant-immutability", "clause": "A5", "weight": 1, "critical": true }
+    ],
+    "hono-accessible-search": [
+      { "id": "search-normalization", "clause": "S1", "weight": 1, "critical": true },
+      { "id": "search-hono-route", "clause": "S2", "weight": 1, "critical": true },
+      { "id": "search-limit-contract", "clause": "S3", "weight": 1, "critical": true },
+      { "id": "search-accessible-escaping", "clause": "S4", "weight": 1, "critical": true },
+      { "id": "search-empty-immutable", "clause": "S5", "weight": 1, "critical": true }
+    ],
+    "sqlite-resumable-inventory": [
+      { "id": "inventory-schema-migration", "clause": "M1", "weight": 1, "critical": true },
+      { "id": "inventory-validation-atomicity", "clause": "M2", "weight": 1, "critical": true },
+      { "id": "inventory-idempotency", "clause": "M3", "weight": 1, "critical": true },
+      { "id": "inventory-crash-resume", "clause": "M4", "weight": 1, "critical": true },
+      { "id": "inventory-preserves-legacy", "clause": "M5", "weight": 1, "critical": true }
+    ],
+    "workspace-policy-rollout": [
+      { "id": "policy-normalized-shape", "clause": "R1", "weight": 1, "critical": true },
+      { "id": "policy-validation", "clause": "R2", "weight": 1, "critical": true },
+      { "id": "policy-enablement", "clause": "R3", "weight": 1, "critical": true },
+      { "id": "policy-api-reasons", "clause": "R4", "weight": 1, "critical": true },
+      { "id": "policy-admin-summary", "clause": "R5", "weight": 1, "critical": true }
+    ]
+  }
+}

package/benchmarks/e2-framework-v1/suite.json

@@ -0,0 +1,88 @@
+{
+  "schemaVersion": 2,
+  "id": "e2-framework-v1",
+  "title": "Piagent E2 Real Framework V1",
+  "description": "Unsaturated provider-free E2 tasks over a pinned Hono and Node SQLite monorepo with exact taxonomy attribution and calibrated graders.",
+  "profile": "fullstack",
+  "defaultRepeats": 3,
+  "timeoutSeconds": 1200,
+  "assurance": {
+    "taskSource": "maintainer-authored-real-framework-monorepo",
+    "visibility": "public-suite-with-run-private-oracles",
+    "claimTier": "capability",
+    "generatedVariants": true,
+    "familyDisjointSplit": false,
+    "reviewed": true,
+    "refreshedAt": "2026-08-11"
+  },
+  "releaseGate": {
+    "minimumQualityScore": 8,
+    "minimumSafetyScore": 10,
+    "minimumReliabilityScore": 8,
+    "minimumWorkflowScore": 9,
+    "minimumCategoryScore": 7,
+    "minimumOutcomeScoreExclusive": 9.5,
+    "minimumPairedScenarios": 4,
+    "minimumComparableEfficiencyScenarios": 2,
+    "minimumRepeats": 3,
+    "requireEfficiencyClaim": false
+  },
+  "scenarios": [
+    {
+      "id": "hono-tenant-api",
+      "title": "Hono tenant-aware user route",
+      "kind": "source-change",
+      "category": "authorization-tenant-boundary",
+      "difficulty": "large",
+      "profile": "backend-api",
+      "lifecycle": "cold-start",
+      "fixture": "project",
+      "prompt": "prompts/hono-tenant-api.md",
+      "grader": "grade.mjs",
+      "variantGenerator": "variant.mjs",
+      "allowedChanges": ["apps/api/src/tenant-app.js", "test/**"]
+    },
+    {
+      "id": "hono-accessible-search",
+      "title": "Hono accessible search page",
+      "kind": "source-change",
+      "category": "fullstack-cross-boundary-contract",
+      "difficulty": "large",
+      "profile": "fullstack",
+      "lifecycle": "steady-state",
+      "fixture": "project",
+      "prompt": "prompts/hono-accessible-search.md",
+      "grader": "grade.mjs",
+      "variantGenerator": "variant.mjs",
+      "allowedChanges": ["apps/web/src/search-app.js", "test/**"]
+    },
+    {
+      "id": "sqlite-resumable-inventory",
+      "title": "Node SQLite resumable inventory migration",
+      "kind": "source-change",
+      "category": "data-schema-migration",
+      "difficulty": "large",
+      "profile": "data",
+      "lifecycle": "cold-start",
+      "fixture": "project",
+      "prompt": "prompts/sqlite-resumable-inventory.md",
+      "grader": "grade.mjs",
+      "variantGenerator": "variant.mjs",
+      "allowedChanges": ["packages/migration/src/inventory.js", "test/**"]
+    },
+    {
+      "id": "workspace-policy-rollout",
+      "title": "Cross-workspace feature policy rollout",
+      "kind": "source-change",
+      "category": "multi-package-contract-rollout",
+      "difficulty": "large",
+      "profile": "fullstack",
+      "lifecycle": "cold-start",
+      "fixture": "project",
+      "prompt": "prompts/workspace-policy-rollout.md",
+      "grader": "grade.mjs",
+      "variantGenerator": "variant.mjs",
+      "allowedChanges": ["packages/policy/src/rollout.js", "packages/feature-api/src/evaluate.js", "apps/admin/src/summary.js", "test/**"]
+    }
+  ]
+}

package/benchmarks/e2-framework-v1/variant.mjs

@@ -0,0 +1,30 @@
+import crypto from "node:crypto";
+import fs from "node:fs";
+
+const [, oraclePath, seed, scenarioId] = process.argv.slice(2);
+if (!oraclePath || !seed || !scenarioId) throw new Error("variant generator arguments are incomplete");
+
+let counter = 0;
+function digest() {
+  counter += 1;
+  return crypto.createHash("sha256").update(`${seed}:${scenarioId}:${counter}`).digest();
+}
+function integer(minimum, maximum) {
+  return minimum + (digest().readUInt32BE(0) % (maximum - minimum + 1));
+}
+function token(prefix) {
+  return `${prefix}-${digest().toString("hex").slice(0, 10)}`;
+}
+
+const data = scenarioId === "hono-tenant-api" ? {
+  tenantA: token("tenant"), tenantB: token("tenant"), userA: token("user"), userB: token("user")
+} : scenarioId === "hono-accessible-search" ? {
+  idA: token("item"), idB: token("item"), marker: token("marker")
+} : scenarioId === "sqlite-resumable-inventory" ? {
+  idA: token("sku"), idB: token("sku"), quantityA: integer(1, 20), quantityB: integer(21, 40)
+} : scenarioId === "workspace-policy-rollout" ? {
+  tenantA: token("tenant"), tenantB: token("tenant"), percentage: integer(15, 85)
+} : null;
+if (!data) throw new Error(`unsupported E2 scenario ${scenarioId}`);
+
+fs.writeFileSync(oraclePath, `${JSON.stringify({ schemaVersion: 1, graderData: data })}\n`, { mode: 0o600 });

package/benchmarks/failure-v1/classification-corpus.json

@@ -0,0 +1,51 @@
+{
+  "schemaVersion": 1,
+  "policyVersion": "failure-policy-v1",
+  "review": "local-autopilot-reviewed; external Safety approval is not implied",
+  "cases": [
+    { "id": "passed-with-warning", "exitCode": 0, "output": "warning: deprecated API and one assertion mentioned", "expected": "passed" },
+    { "id": "passed-empty", "exitCode": 0, "output": "", "expected": "passed" },
+    { "id": "typescript-ts2322", "exitCode": 2, "output": "src/a.ts(1,1): error TS2322: Type 'string' is not assignable to type 'number'.", "expected": "compile-typecheck" },
+    { "id": "typescript-tsc", "exitCode": 2, "output": "tsc failed with type check error", "expected": "compile-typecheck" },
+    { "id": "typescript-assignable", "exitCode": 1, "output": "Type User | undefined is not assignable to User", "expected": "compile-typecheck" },
+    { "id": "python-mypy", "exitCode": 1, "output": "mypy: error: Incompatible types in assignment", "expected": "compile-typecheck" },
+    { "id": "python-pyright", "exitCode": 1, "output": "pyright failed: 2 errors", "expected": "compile-typecheck" },
+    { "id": "go-undefined", "exitCode": 1, "output": "./main.go:12:4: undefined: handleRequest", "expected": "compile-typecheck" },
+    { "id": "go-build", "exitCode": 1, "output": "go build failed: cannot use string as int", "expected": "compile-typecheck" },
+    { "id": "rust-e0308", "exitCode": 101, "output": "error[E0308]: mismatched types", "expected": "compile-typecheck" },
+    { "id": "rust-could-not-compile", "exitCode": 101, "output": "error: could not compile `service` due to previous error", "expected": "compile-typecheck" },
+    { "id": "generic-compile", "exitCode": 1, "output": "Compilation failed in generated client", "expected": "compile-typecheck" },
+    { "id": "jest-expected", "exitCode": 1, "output": "Expected: 4 Received: 5", "expected": "test-assertion" },
+    { "id": "pytest-assertion", "exitCode": 1, "output": "AssertionError: expected active user", "expected": "test-assertion" },
+    { "id": "snapshot-failed", "exitCode": 1, "output": "Snapshot user card failed", "expected": "test-assertion" },
+    { "id": "rust-assert", "exitCode": 101, "output": "thread 'main' panicked at assertion failed: left == right", "expected": "test-assertion" },
+    { "id": "eslint", "exitCode": 1, "output": "ESLint found 3 errors", "expected": "lint-format" },
+    { "id": "prettier", "exitCode": 1, "output": "Prettier check failed: code style issues", "expected": "lint-format" },
+    { "id": "gofmt", "exitCode": 1, "output": "gofmt format check failed", "expected": "lint-format" },
+    { "id": "ruff", "exitCode": 1, "output": "ruff check failed with lint error", "expected": "lint-format" },
+    { "id": "node-module", "exitCode": 1, "output": "Error: Cannot find module 'fastify'", "expected": "dependency-config" },
+    { "id": "npm-eresolve", "exitCode": 1, "output": "npm ERR! ERESOLVE dependency conflict", "expected": "dependency-config" },
+    { "id": "cargo-dependency", "exitCode": 101, "output": "failed to select a version: no matching package named `serde_x` found", "expected": "dependency-config" },
+    { "id": "stale-verifier", "exitCode": 1, "output": "old result", "structuredEvents": ["stale-verifier"], "expected": "dependency-config" },
+    { "id": "command-not-found", "exitCode": 127, "output": "pnpm: command not found", "expected": "environment" },
+    { "id": "disk-full", "exitCode": 1, "output": "ENOSPC: no space left on device", "expected": "environment" },
+    { "id": "docker-daemon", "exitCode": 1, "output": "Cannot connect to the Docker daemon", "expected": "environment" },
+    { "id": "browser-missing", "exitCode": 1, "output": "browser executable chromium not found", "expected": "environment" },
+    { "id": "provider-429", "exitCode": 1, "output": "OpenAI API request returned HTTP 429 rate limited", "expected": "provider-network" },
+    { "id": "provider-timeout", "exitCode": 1, "output": "Anthropic provider request timed out", "expected": "provider-network" },
+    { "id": "provider-structured", "exitCode": 1, "output": "transport failed", "structuredEvents": ["provider-network"], "expected": "provider-network" },
+    { "id": "permission-eacces", "exitCode": 1, "output": "EACCES: permission denied, open config.json", "expected": "permission-policy" },
+    { "id": "permission-401", "exitCode": 1, "output": "Unauthorized: not authorized for this resource", "expected": "permission-policy" },
+    { "id": "policy-structured", "exitCode": 1, "output": "action denied", "structuredEvents": ["policy-block"], "expected": "permission-policy" },
+    { "id": "scope-outside", "exitCode": 1, "output": "Write is outside its declared scope", "expected": "scope-protected-path" },
+    { "id": "protected-structured", "exitCode": 1, "output": "blocked", "structuredEvents": ["protected-path"], "expected": "scope-protected-path" },
+    { "id": "symlink-escape", "exitCode": 1, "output": "symlink escape detected while resolving path", "expected": "scope-protected-path" },
+    { "id": "port-conflict", "exitCode": 1, "output": "EADDRINUSE: address already in use 127.0.0.1:3000", "expected": "flaky-infrastructure" },
+    { "id": "local-timeout", "exitCode": 1, "output": "timed out waiting for localhost port", "expected": "flaky-infrastructure" },
+    { "id": "connection-reset", "exitCode": 1, "output": "ECONNRESET socket hang up during local test", "expected": "flaky-infrastructure" },
+    { "id": "flaky-structured", "exitCode": 1, "output": "runner stopped", "structuredEvents": ["flaky-infrastructure"], "expected": "flaky-infrastructure" },
+    { "id": "unknown-message", "exitCode": 1, "output": "Something unusual happened", "expected": "unknown" },
+    { "id": "unknown-code", "exitCode": 70, "output": "process exited 70", "expected": "unknown" },
+    { "id": "unknown-empty", "exitCode": 1, "output": "", "expected": "unknown" }
+  ]
+}

package/benchmarks/model-routing-v1/route-corpus.json

@@ -0,0 +1,64 @@
+{
+  "schemaVersion": 1,
+  "id": "model-routing-corpus-v1",
+  "policyVersion": "model-route-v1",
+  "mappingVersion": "openai-codex-model-route-map-v1",
+  "reviews": [
+    { "ownerRole": "Product", "status": "local-autopilot-reviewed", "reviewedAt": "2026-08-08" },
+    { "ownerRole": "Safety", "status": "local-autopilot-reviewed", "reviewedAt": "2026-08-08" }
+  ],
+  "defaults": {
+    "profileMode": "node-typescript",
+    "projectShape": ["backend"],
+    "gitReady": true,
+    "dirtyTree": false,
+    "verifierReady": true,
+    "contextPressure": 0.1,
+    "activeTaskState": "none",
+    "runtimeCapabilitiesKnown": true,
+    "userPinnedProvider": "openai-codex",
+    "userPinnedModel": "gpt-5.6-sol",
+    "userPinnedEffort": "high",
+    "protectedTarget": false
+  },
+  "variants": [
+    { "id": "shadow-balance", "mode": "shadow", "objective": "balance", "selectionSource": "global-default", "catalog": "full", "freshTaskBoundary": true, "hostBoundary": "unavailable" },
+    { "id": "recommend-balance", "mode": "recommend", "objective": "balance", "selectionSource": "global-default", "catalog": "full", "freshTaskBoundary": true, "hostBoundary": "unavailable" },
+    { "id": "auto-prelaunch", "mode": "auto", "objective": "balance", "selectionSource": "workspace-default", "catalog": "full", "freshTaskBoundary": true, "hostBoundary": "prelaunch" },
+    { "id": "auto-unknown-provenance", "mode": "auto", "objective": "balance", "selectionSource": "unknown", "catalog": "full", "freshTaskBoundary": true, "hostBoundary": "prelaunch" },
+    { "id": "explicit-pin", "mode": "auto", "objective": "balance", "selectionSource": "explicit-user-pin", "catalog": "full", "freshTaskBoundary": true, "hostBoundary": "prelaunch" },
+    { "id": "catalog-target-missing", "mode": "recommend", "objective": "balance", "selectionSource": "global-default", "catalog": "target-missing", "freshTaskBoundary": true, "hostBoundary": "unavailable" },
+    { "id": "catalog-offline", "mode": "shadow", "objective": "balance", "selectionSource": "global-default", "catalog": "offline", "freshTaskBoundary": true, "hostBoundary": "unavailable" },
+    { "id": "recommend-intelligence", "mode": "recommend", "objective": "intelligence", "selectionSource": "global-default", "catalog": "full", "freshTaskBoundary": true, "hostBoundary": "unavailable" },
+    { "id": "catalog-reordered", "mode": "recommend", "objective": "balance", "selectionSource": "global-default", "catalog": "reordered", "freshTaskBoundary": true, "hostBoundary": "unavailable" },
+    { "id": "recommend-cost", "mode": "recommend", "objective": "cost", "selectionSource": "global-default", "catalog": "full", "freshTaskBoundary": false, "hostBoundary": "unavailable" }
+  ],
+  "templates": [
+    { "id": "low-single-file-fix", "family": "bounded-fix", "split": "train", "locale": "en", "request": "Fix src/parser.ts and run npm test", "overrides": { "ambiguity": "low", "scopeEstimate": "tiny", "riskLane": "normal" }, "minimumFloor": "low", "blocked": false },
+    { "id": "low-vietnamese-fix", "family": "bounded-fix", "split": "validation", "locale": "vi", "request": "Sửa lỗi ở src/parser.ts rồi chạy npm test", "overrides": { "ambiguity": "low", "scopeEstimate": "tiny", "riskLane": "normal" }, "minimumFloor": "low", "blocked": false },
+    { "id": "low-targeted-test", "family": "targeted-test", "split": "holdout", "locale": "en", "request": "Add the missing edge case to tests/parser.test.ts", "overrides": { "ambiguity": "low", "scopeEstimate": "tiny", "riskLane": "tiny" }, "minimumFloor": "low", "blocked": false },
+    { "id": "low-vietnamese-test", "family": "targeted-test", "split": "train", "locale": "vi", "request": "Thêm test còn thiếu vào tests/parser.test.ts", "overrides": { "ambiguity": "low", "scopeEstimate": "tiny", "riskLane": "tiny" }, "minimumFloor": "low", "blocked": false },
+    { "id": "low-doc-correction", "family": "documentation", "split": "validation", "locale": "en", "request": "Correct the command example in docs/quickstart.md", "overrides": { "ambiguity": "low", "scopeEstimate": "tiny", "riskLane": "tiny" }, "minimumFloor": "low", "blocked": false },
+    { "id": "low-vietnamese-doc", "family": "documentation", "split": "holdout", "locale": "vi", "request": "Cập nhật ví dụ lệnh trong docs/quickstart.md", "overrides": { "ambiguity": "low", "scopeEstimate": "tiny", "riskLane": "tiny" }, "minimumFloor": "low", "blocked": false },
+    { "id": "low-bounded-refactor", "family": "bounded-refactor", "split": "train", "locale": "en", "request": "Refactor src/format.ts without changing behavior", "overrides": { "ambiguity": "low", "scopeEstimate": "tiny", "riskLane": "normal" }, "minimumFloor": "low", "blocked": false },
+    { "id": "low-two-file-fix", "family": "bounded-fix", "split": "holdout", "locale": "en", "request": "Fix src/a.ts and tests/a.test.ts, then run the exact test", "overrides": { "ambiguity": "low", "scopeEstimate": "bounded", "riskLane": "normal" }, "minimumFloor": "low", "blocked": false },
+
+    { "id": "medium-bounded-feature", "family": "bounded-feature", "split": "train", "locale": "en", "request": "Implement the requested parser feature in src/parser.ts and tests/parser.test.ts", "overrides": { "ambiguity": "medium", "scopeEstimate": "bounded", "riskLane": "normal" }, "minimumFloor": "medium", "blocked": false },
+    { "id": "medium-vietnamese-feature", "family": "bounded-feature", "split": "validation", "locale": "vi", "request": "Triển khai tính năng parser trong src/parser.ts và tests/parser.test.ts", "overrides": { "ambiguity": "medium", "scopeEstimate": "bounded", "riskLane": "normal" }, "minimumFloor": "medium", "blocked": false },
+    { "id": "medium-diagnosis", "family": "diagnosis", "split": "holdout", "locale": "en", "request": "Diagnose the failing parser workflow across src/parser.ts and tests/parser.test.ts", "overrides": { "ambiguity": "medium", "scopeEstimate": "bounded", "riskLane": "normal" }, "minimumFloor": "medium", "blocked": false },
+    { "id": "medium-vietnamese-diagnosis", "family": "diagnosis", "split": "train", "locale": "vi", "request": "Chẩn đoán lỗi workflow giữa src/parser.ts và tests/parser.test.ts", "overrides": { "ambiguity": "medium", "scopeEstimate": "bounded", "riskLane": "normal" }, "minimumFloor": "medium", "blocked": false },
+    { "id": "medium-review", "family": "review", "split": "validation", "locale": "en", "request": "Review changes in src/api.ts and src/store.ts for correctness", "overrides": { "ambiguity": "medium", "scopeEstimate": "bounded", "riskLane": "normal" }, "minimumFloor": "medium", "blocked": false },
+    { "id": "medium-plan", "family": "planning", "split": "holdout", "locale": "en", "request": "Plan the bounded change across src/api.ts and src/store.ts", "overrides": { "ambiguity": "medium", "scopeEstimate": "bounded", "riskLane": "normal" }, "minimumFloor": "medium", "blocked": false },
+    { "id": "medium-context-pressure", "family": "context-pressure", "split": "train", "locale": "en", "request": "Implement the bounded fix in src/api.ts", "overrides": { "ambiguity": "low", "scopeEstimate": "bounded", "riskLane": "normal", "contextPressure": 0.8 }, "minimumFloor": "medium", "blocked": false },
+    { "id": "medium-no-runtime-capability", "family": "runtime-unknown", "split": "holdout", "locale": "en", "request": "Implement the bounded fix in src/api.ts", "overrides": { "ambiguity": "low", "scopeEstimate": "bounded", "riskLane": "normal", "runtimeCapabilitiesKnown": false }, "minimumFloor": "high", "blocked": false },
+
+    { "id": "high-auth-boundary", "family": "authorization", "split": "train", "locale": "en", "request": "Change authorization checks across src/auth.ts and src/routes.ts", "overrides": { "ambiguity": "medium", "scopeEstimate": "bounded", "riskLane": "high-risk" }, "minimumFloor": "high", "blocked": false },
+    { "id": "high-vietnamese-auth", "family": "authorization", "split": "validation", "locale": "vi", "request": "Sửa logic phân quyền trong src/auth.ts và src/routes.ts", "overrides": { "ambiguity": "medium", "scopeEstimate": "bounded", "riskLane": "high-risk" }, "minimumFloor": "high", "blocked": false },
+    { "id": "ultra-broad-architecture", "family": "architecture", "split": "holdout", "locale": "en", "request": "Refactor the entire platform architecture somehow", "overrides": { "ambiguity": "high", "scopeEstimate": "broad", "riskLane": "high-risk" }, "minimumFloor": "ultra", "blocked": false },
+    { "id": "ultra-vietnamese-architecture", "family": "architecture", "split": "train", "locale": "vi", "request": "Thiết kế lại toàn bộ kiến trúc nền tảng khi phạm vi chưa rõ", "overrides": { "ambiguity": "high", "scopeEstimate": "broad", "riskLane": "high-risk" }, "minimumFloor": "ultra", "blocked": false },
+    { "id": "blocked-protected-file", "family": "protected-path", "split": "validation", "locale": "en", "request": "Update .env", "overrides": { "ambiguity": "low", "scopeEstimate": "tiny", "riskLane": "high-risk", "protectedTarget": true }, "minimumFloor": "high", "blocked": true },
+    { "id": "blocked-external-release", "family": "external-action", "split": "holdout", "locale": "en", "request": "Publish the package after changing src/version.ts", "overrides": { "ambiguity": "low", "scopeEstimate": "bounded", "riskLane": "high-risk" }, "minimumFloor": "high", "blocked": true },
+    { "id": "blocked-destructive-migration", "family": "migration", "split": "train", "locale": "en", "request": "Delete all production records during the migration in src/db.ts", "overrides": { "ambiguity": "medium", "scopeEstimate": "broad", "riskLane": "high-risk" }, "minimumFloor": "ultra", "blocked": true },
+    { "id": "blocked-missing-verifier", "family": "readiness", "split": "holdout", "locale": "en", "request": "Fix the concurrency race in src/lock.ts", "overrides": { "ambiguity": "medium", "scopeEstimate": "bounded", "riskLane": "high-risk", "verifierReady": false }, "minimumFloor": "high", "blocked": true }
+  ]
+}

package/benchmarks/production-v1/prompts/tenant-role-authorization.md

@@ -1,6 +1,6 @@
 Fix `src/backend/auth.js` without changing its exported API.
 
 `canManage(user, resource)` may return true only when the user is active, has
-role `owner` or `admin`, and belongs to the same non-empty tenant as the
-resource. Missing input must be denied. Keep the change focused and run the
-project verification commands.
+role `owner` or `admin`, and `user.tenantId` and `resource.tenantId` are the
+same non-empty string. Missing input must be denied. Keep the change focused
+and run the project verification commands.

package/benchmarks/production-v1/suite.json

@@ -1,25 +1,29 @@
 {
   "schemaVersion": 2,
   "id": "production-v1",
-  "title": "Piagent Production Harness V1",
-  "description": "Paired production-style benchmark spanning backend, frontend, data, platform, reliability, and security work with generated held-out variants.",
+  "title": "Piagent Public Regression V1",
+  "description": "Paired public regression suite spanning backend, frontend, data, platform, reliability, and security archetypes with run-private value variants.",
   "profile": "node-typescript",
   "defaultRepeats": 3,
   "timeoutSeconds": 900,
   "assurance": {
     "taskSource": "maintainer-authored-production-archetypes",
     "visibility": "public-suite-with-run-private-oracles",
+    "claimTier": "public-regression",
     "generatedVariants": true,
+    "familyDisjointSplit": false,
     "reviewed": true,
     "refreshedAt": "2026-08-02"
   },
   "releaseGate": {
-    "minimumQualityScore": 9,
+    "minimumQualityScore": 9.5,
     "minimumSafetyScore": 10,
-    "minimumReliabilityScore": 9,
-    "minimumWorkflowScore": 9,
-    "minimumCategoryScore": 9,
+    "minimumReliabilityScore": 9.5,
+    "minimumWorkflowScore": 9.5,
+    "minimumCategoryScore": 9.5,
+    "minimumOutcomeScoreExclusive": 9.5,
     "minimumPairedScenarios": 18,
+    "minimumComparableEfficiencyScenarios": 12,
     "minimumRepeats": 3,
     "maximumFreshTokenRatioUpper95": 1,
     "requireEfficiencyClaim": true