@piagent/platform 1.5.1 → 1.5.3

package/docs-site/en/permissions.html

@@ -30,7 +30,7 @@
         <span>Pi Agent Platform</span>
       </a>
       <div class="top-actions">
-        <span class="pill"><span class="pill-dot"></span>v1.5.1 docs · EN</span>
+        <span class="pill"><span class="pill-dot"></span>v1.5.3 docs · EN</span>
         <nav class="language-switch" aria-label="Language">
           <a class="language-option" href="/permissions" hreflang="vi" lang="vi">VI</a>
           <a class="language-option current" href="/en/permissions" hreflang="en" lang="en" aria-current="true">EN</a>

package/docs-site/en/profiles.html

@@ -30,7 +30,7 @@
         <span>Pi Agent Platform</span>
       </a>
       <div class="top-actions">
-        <span class="pill"><span class="pill-dot"></span>v1.5.1 docs · EN</span>
+        <span class="pill"><span class="pill-dot"></span>v1.5.3 docs · EN</span>
         <nav class="language-switch" aria-label="Language">
           <a class="language-option" href="/profiles" hreflang="vi" lang="vi">VI</a>
           <a class="language-option current" href="/en/profiles" hreflang="en" lang="en" aria-current="true">EN</a>

package/docs-site/en/quickstart.html

@@ -30,7 +30,7 @@
         <span>Pi Agent Platform</span>
       </a>
       <div class="top-actions">
-        <span class="pill"><span class="pill-dot"></span>v1.5.1 docs · EN</span>
+        <span class="pill"><span class="pill-dot"></span>v1.5.3 docs · EN</span>
         <nav class="language-switch" aria-label="Language">
           <a class="language-option" href="/quickstart" hreflang="vi" lang="vi">VI</a>
           <a class="language-option current" href="/en/quickstart" hreflang="en" lang="en" aria-current="true">EN</a>

package/docs-site/en/runtime.html

@@ -30,7 +30,7 @@
         <span>Pi Agent Platform</span>
       </a>
       <div class="top-actions">
-        <span class="pill"><span class="pill-dot"></span>v1.5.1 docs · EN</span>
+        <span class="pill"><span class="pill-dot"></span>v1.5.3 docs · EN</span>
         <nav class="language-switch" aria-label="Language">
           <a class="language-option" href="/runtime" hreflang="vi" lang="vi">VI</a>
           <a class="language-option current" href="/en/runtime" hreflang="en" lang="en" aria-current="true">EN</a>

package/docs-site/en/scope.html

@@ -30,7 +30,7 @@
         <span>Pi Agent Platform</span>
       </a>
       <div class="top-actions">
-        <span class="pill"><span class="pill-dot"></span>v1.5.1 docs · EN</span>
+        <span class="pill"><span class="pill-dot"></span>v1.5.3 docs · EN</span>
         <nav class="language-switch" aria-label="Language">
           <a class="language-option" href="/scope" hreflang="vi" lang="vi">VI</a>
           <a class="language-option current" href="/en/scope" hreflang="en" lang="en" aria-current="true">EN</a>

package/docs-site/en/security.html

@@ -30,7 +30,7 @@
         <span>Pi Agent Platform</span>
       </a>
       <div class="top-actions">
-        <span class="pill"><span class="pill-dot"></span>v1.5.1 docs · EN</span>
+        <span class="pill"><span class="pill-dot"></span>v1.5.3 docs · EN</span>
         <nav class="language-switch" aria-label="Language">
           <a class="language-option" href="/security" hreflang="vi" lang="vi">VI</a>
           <a class="language-option current" href="/en/security" hreflang="en" lang="en" aria-current="true">EN</a>

package/docs-site/en/team.html

@@ -30,7 +30,7 @@
         <span>Pi Agent Platform</span>
       </a>
       <div class="top-actions">
-        <span class="pill"><span class="pill-dot"></span>v1.5.1 docs · EN</span>
+        <span class="pill"><span class="pill-dot"></span>v1.5.3 docs · EN</span>
         <nav class="language-switch" aria-label="Language">
           <a class="language-option" href="/team" hreflang="vi" lang="vi">VI</a>
           <a class="language-option current" href="/en/team" hreflang="en" lang="en" aria-current="true">EN</a>

package/docs-site/en/whats-new.html

@@ -30,7 +30,7 @@
         <span>Pi Agent Platform</span>
       </a>
       <div class="top-actions">
-        <span class="pill"><span class="pill-dot"></span>v1.5.1 docs · EN</span>
+        <span class="pill"><span class="pill-dot"></span>v1.5.3 docs · EN</span>
         <nav class="language-switch" aria-label="Language">
           <a class="language-option" href="/whats-new" hreflang="vi" lang="vi">VI</a>
           <a class="language-option current" href="/en/whats-new" hreflang="en" lang="en" aria-current="true">EN</a>

package/docs-site/en/workflows.html

@@ -30,7 +30,7 @@
         <span>Pi Agent Platform</span>
       </a>
       <div class="top-actions">
-        <span class="pill"><span class="pill-dot"></span>v1.5.1 docs · EN</span>
+        <span class="pill"><span class="pill-dot"></span>v1.5.3 docs · EN</span>
         <nav class="language-switch" aria-label="Language">
           <a class="language-option" href="/workflows" hreflang="vi" lang="vi">VI</a>
           <a class="language-option current" href="/en/workflows" hreflang="en" lang="en" aria-current="true">EN</a>

package/docs-site/index.html

@@ -30,7 +30,7 @@
         <span>Pi Agent Platform</span>
       </a>
       <div class="top-actions">
-        <span class="pill"><span class="pill-dot"></span>v1.5.1 docs · VI</span>
+        <span class="pill"><span class="pill-dot"></span>v1.5.3 docs · VI</span>
         <nav class="language-switch" aria-label="Ngôn ngữ">
           <a class="language-option current" href="/" hreflang="vi" lang="vi" aria-current="true">VI</a>
           <a class="language-option" href="/en/" hreflang="en" lang="en">EN</a>

package/docs-site/mcp-approval.html

@@ -30,7 +30,7 @@
         <span>Pi Agent Platform</span>
       </a>
       <div class="top-actions">
-        <span class="pill"><span class="pill-dot"></span>v1.5.1 docs · VI</span>
+        <span class="pill"><span class="pill-dot"></span>v1.5.3 docs · VI</span>
         <nav class="language-switch" aria-label="Ngôn ngữ">
           <a class="language-option current" href="/mcp-approval" hreflang="vi" lang="vi" aria-current="true">VI</a>
           <a class="language-option" href="/en/mcp-approval" hreflang="en" lang="en">EN</a>

package/docs-site/mcp-auth.html

@@ -30,7 +30,7 @@
         <span>Pi Agent Platform</span>
       </a>
       <div class="top-actions">
-        <span class="pill"><span class="pill-dot"></span>v1.5.1 docs · VI</span>
+        <span class="pill"><span class="pill-dot"></span>v1.5.3 docs · VI</span>
         <nav class="language-switch" aria-label="Ngôn ngữ">
           <a class="language-option current" href="/mcp-auth" hreflang="vi" lang="vi" aria-current="true">VI</a>
           <a class="language-option" href="/en/mcp-auth" hreflang="en" lang="en">EN</a>

package/docs-site/mcp-servers.html

@@ -30,7 +30,7 @@
         <span>Pi Agent Platform</span>
       </a>
       <div class="top-actions">
-        <span class="pill"><span class="pill-dot"></span>v1.5.1 docs · VI</span>
+        <span class="pill"><span class="pill-dot"></span>v1.5.3 docs · VI</span>
         <nav class="language-switch" aria-label="Ngôn ngữ">
           <a class="language-option current" href="/mcp-servers" hreflang="vi" lang="vi" aria-current="true">VI</a>
           <a class="language-option" href="/en/mcp-servers" hreflang="en" lang="en">EN</a>

package/docs-site/mcp.html

@@ -30,7 +30,7 @@
         <span>Pi Agent Platform</span>
       </a>
       <div class="top-actions">
-        <span class="pill"><span class="pill-dot"></span>v1.5.1 docs · VI</span>
+        <span class="pill"><span class="pill-dot"></span>v1.5.3 docs · VI</span>
         <nav class="language-switch" aria-label="Ngôn ngữ">
           <a class="language-option current" href="/mcp" hreflang="vi" lang="vi" aria-current="true">VI</a>
           <a class="language-option" href="/en/mcp" hreflang="en" lang="en">EN</a>

package/docs-site/permissions.html

@@ -30,7 +30,7 @@
         <span>Pi Agent Platform</span>
       </a>
       <div class="top-actions">
-        <span class="pill"><span class="pill-dot"></span>v1.5.1 docs · VI</span>
+        <span class="pill"><span class="pill-dot"></span>v1.5.3 docs · VI</span>
         <nav class="language-switch" aria-label="Ngôn ngữ">
           <a class="language-option current" href="/permissions" hreflang="vi" lang="vi" aria-current="true">VI</a>
           <a class="language-option" href="/en/permissions" hreflang="en" lang="en">EN</a>

package/docs-site/profiles.html

@@ -30,7 +30,7 @@
         <span>Pi Agent Platform</span>
       </a>
       <div class="top-actions">
-        <span class="pill"><span class="pill-dot"></span>v1.5.1 docs · VI</span>
+        <span class="pill"><span class="pill-dot"></span>v1.5.3 docs · VI</span>
         <nav class="language-switch" aria-label="Ngôn ngữ">
           <a class="language-option current" href="/profiles" hreflang="vi" lang="vi" aria-current="true">VI</a>
           <a class="language-option" href="/en/profiles" hreflang="en" lang="en">EN</a>

package/docs-site/quickstart.html

@@ -30,7 +30,7 @@
         <span>Pi Agent Platform</span>
       </a>
       <div class="top-actions">
-        <span class="pill"><span class="pill-dot"></span>v1.5.1 docs · VI</span>
+        <span class="pill"><span class="pill-dot"></span>v1.5.3 docs · VI</span>
         <nav class="language-switch" aria-label="Ngôn ngữ">
           <a class="language-option current" href="/quickstart" hreflang="vi" lang="vi" aria-current="true">VI</a>
           <a class="language-option" href="/en/quickstart" hreflang="en" lang="en">EN</a>

package/docs-site/runtime.html

@@ -30,7 +30,7 @@
         <span>Pi Agent Platform</span>
       </a>
       <div class="top-actions">
-        <span class="pill"><span class="pill-dot"></span>v1.5.1 docs · VI</span>
+        <span class="pill"><span class="pill-dot"></span>v1.5.3 docs · VI</span>
         <nav class="language-switch" aria-label="Ngôn ngữ">
           <a class="language-option current" href="/runtime" hreflang="vi" lang="vi" aria-current="true">VI</a>
           <a class="language-option" href="/en/runtime" hreflang="en" lang="en">EN</a>

package/docs-site/scope.html

@@ -30,7 +30,7 @@
         <span>Pi Agent Platform</span>
       </a>
       <div class="top-actions">
-        <span class="pill"><span class="pill-dot"></span>v1.5.1 docs · VI</span>
+        <span class="pill"><span class="pill-dot"></span>v1.5.3 docs · VI</span>
         <nav class="language-switch" aria-label="Ngôn ngữ">
           <a class="language-option current" href="/scope" hreflang="vi" lang="vi" aria-current="true">VI</a>
           <a class="language-option" href="/en/scope" hreflang="en" lang="en">EN</a>

package/docs-site/security.html

@@ -30,7 +30,7 @@
         <span>Pi Agent Platform</span>
       </a>
       <div class="top-actions">
-        <span class="pill"><span class="pill-dot"></span>v1.5.1 docs · VI</span>
+        <span class="pill"><span class="pill-dot"></span>v1.5.3 docs · VI</span>
         <nav class="language-switch" aria-label="Ngôn ngữ">
           <a class="language-option current" href="/security" hreflang="vi" lang="vi" aria-current="true">VI</a>
           <a class="language-option" href="/en/security" hreflang="en" lang="en">EN</a>

package/docs-site/team.html

@@ -30,7 +30,7 @@
         <span>Pi Agent Platform</span>
       </a>
       <div class="top-actions">
-        <span class="pill"><span class="pill-dot"></span>v1.5.1 docs · VI</span>
+        <span class="pill"><span class="pill-dot"></span>v1.5.3 docs · VI</span>
         <nav class="language-switch" aria-label="Ngôn ngữ">
           <a class="language-option current" href="/team" hreflang="vi" lang="vi" aria-current="true">VI</a>
           <a class="language-option" href="/en/team" hreflang="en" lang="en">EN</a>

package/docs-site/whats-new.html

@@ -30,7 +30,7 @@
         <span>Pi Agent Platform</span>
       </a>
       <div class="top-actions">
-        <span class="pill"><span class="pill-dot"></span>v1.5.1 docs · VI</span>
+        <span class="pill"><span class="pill-dot"></span>v1.5.3 docs · VI</span>
         <nav class="language-switch" aria-label="Ngôn ngữ">
           <a class="language-option current" href="/whats-new" hreflang="vi" lang="vi" aria-current="true">VI</a>
           <a class="language-option" href="/en/whats-new" hreflang="en" lang="en">EN</a>

package/docs-site/workflows.html

@@ -30,7 +30,7 @@
         <span>Pi Agent Platform</span>
       </a>
       <div class="top-actions">
-        <span class="pill"><span class="pill-dot"></span>v1.5.1 docs · VI</span>
+        <span class="pill"><span class="pill-dot"></span>v1.5.3 docs · VI</span>
         <nav class="language-switch" aria-label="Ngôn ngữ">
           <a class="language-option current" href="/workflows" hreflang="vi" lang="vi" aria-current="true">VI</a>
           <a class="language-option" href="/en/workflows" hreflang="en" lang="en">EN</a>

package/evals/fixtures/benchmark-assurance-evidence.invalid.json

@@ -28,8 +28,8 @@
     "expiresAt": "2099-09-11T00:00:00.000Z"
   },
   "disjointness": {
-    "taxonomyDigest": "b3fa32c7008ce7b60de4ffebd6853da1601effae8b422f1cfa68810e4ea2150a",
-    "publicExposureDigest": "b91b86607d1499c13e87d577257e14f29fa347f50b1bd9eccab5e5bb64e1549d",
+    "taxonomyDigest": "d42458d04861a60488a5f98532936db71d43dcccc0298d75b09c2255f2922cb1",
+    "publicExposureDigest": "b7bdf8d76f66982410f047b5138ca7b974f9b441115761a7e81817753bfa41b3",
     "reportDigest": "5555555555555555555555555555555555555555555555555555555555555555",
     "familyDisjoint": true,
     "repositoryDisjoint": false,

package/evals/fixtures/benchmark-assurance-evidence.valid.json

@@ -28,8 +28,8 @@
     "expiresAt": "2099-09-11T00:00:00.000Z"
   },
   "disjointness": {
-    "taxonomyDigest": "b3fa32c7008ce7b60de4ffebd6853da1601effae8b422f1cfa68810e4ea2150a",
-    "publicExposureDigest": "b91b86607d1499c13e87d577257e14f29fa347f50b1bd9eccab5e5bb64e1549d",
+    "taxonomyDigest": "d42458d04861a60488a5f98532936db71d43dcccc0298d75b09c2255f2922cb1",
+    "publicExposureDigest": "b7bdf8d76f66982410f047b5138ca7b974f9b441115761a7e81817753bfa41b3",
     "reportDigest": "5555555555555555555555555555555555555555555555555555555555555555",
     "familyDisjoint": true,
     "repositoryDisjoint": true,

package/evals/fixtures/piagent-webui/runtime-command-v1.invalid.json

@@ -0,0 +1,12 @@
+{
+  "schemaVersion": 1,
+  "version": "piagent-runtime-command-v1",
+  "messageType": "command",
+  "requestId": "runtime_request_01",
+  "sessionRef": "session_ref_01",
+  "expectedSessionRevision": "session_revision_01",
+  "action": "context.rebuild",
+  "argument": null,
+  "confirmed": false,
+  "rawCommand": "/context rebuild"
+}

package/evals/fixtures/piagent-webui/runtime-command-v1.valid.json

@@ -0,0 +1,11 @@
+{
+  "schemaVersion": 1,
+  "version": "piagent-runtime-command-v1",
+  "messageType": "command",
+  "requestId": "runtime_request_01",
+  "sessionRef": "session_ref_01",
+  "expectedSessionRevision": "session_revision_01",
+  "action": "context.search",
+  "argument": "runtime command parity",
+  "confirmed": false
+}

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

@@ -17,8 +17,8 @@
     { "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" }
+    { "id": "e3-public-exposure", "path": "evals/private-holdout-v1/public-exposure.v1.json", "sha256": "b7bdf8d76f66982410f047b5138ca7b974f9b441115761a7e81817753bfa41b3" },
+    { "id": "taxonomy", "path": "evals/real-task-taxonomy.v1.json", "sha256": "d42458d04861a60488a5f98532936db71d43dcccc0298d75b09c2255f2922cb1" }
   ],
   "deterministicTests": {
     "e0": ["tests/golden-enforcement.test.mjs", "tests/benchmark-core.test.mjs"],

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

@@ -37,7 +37,7 @@
     },
     {
       "path": "evals/fs5-pilot-protocol.v1.json",
-      "sha256": "d27b5adb9a37a4e7b998f9bd6dafc4c6ed0c8d9da1a013a0c734b8ecc5ff6079"
+      "sha256": "392ec00603a0ef997d18d7327e360dddfe301a748f2395530bffe214c59c9219"
     }
   ],
   "localReproducer": {

package/evals/fs5-pilot-protocol.v1.json

@@ -23,7 +23,7 @@
     {
       "id": "canary-suite",
       "path": "benchmarks/capability-v1/suite.json",
-      "sha256": "378133ad8d10a47e580ecb62b1c0221d562d81a49c8992f3ab8caa46f5980ddc"
+      "sha256": "29bb79e96b047d79056626d747baa0b3a6ca9d4da7ad80b9668c0085e5581706"
     },
     {
       "id": "pilot-suite",

package/evals/fs5-pilot-protocol.v4.json

@@ -16,7 +16,7 @@
   },
   "artifactBindings": [
     { "id": "authority-manifest", "path": "packages/piagent-core/policy/authority-manifest.v1.json", "sha256": "911dc026f1a860f593a6850f7348311f6adac83a8f11627abb23654960b605d0" },
-    { "id": "canary-suite", "path": "benchmarks/capability-v1/suite.json", "sha256": "378133ad8d10a47e580ecb62b1c0221d562d81a49c8992f3ab8caa46f5980ddc" },
+    { "id": "canary-suite", "path": "benchmarks/capability-v1/suite.json", "sha256": "29bb79e96b047d79056626d747baa0b3a6ca9d4da7ad80b9668c0085e5581706" },
     { "id": "pilot-suite", "path": "benchmarks/production-v1/suite.json", "sha256": "8d03df7c08eee5debbabdb3490beb5d45b021a392b6da77df257e6d94c346d58" },
     { "id": "task-contract-schema", "path": "schemas/task-contract.schema.json", "sha256": "432a253af865f5c79527b3c7e5e65fc3b6cfea43d19f9e915f3bda792a358bb4" },
     { "id": "assurance-schema", "path": "schemas/benchmark-assurance-evidence.schema.json", "sha256": "5d937632d2cfff34dbbb781dd367bb78feb64c36d571c36ca15f2cc40c576af8" },

package/evals/fs5-pilot-protocol.v5.json

@@ -16,7 +16,7 @@
   },
   "artifactBindings": [
     { "id": "authority-manifest", "path": "packages/piagent-core/policy/authority-manifest.v1.json", "sha256": "911dc026f1a860f593a6850f7348311f6adac83a8f11627abb23654960b605d0" },
-    { "id": "canary-suite", "path": "benchmarks/capability-v1/suite.json", "sha256": "378133ad8d10a47e580ecb62b1c0221d562d81a49c8992f3ab8caa46f5980ddc" },
+    { "id": "canary-suite", "path": "benchmarks/capability-v1/suite.json", "sha256": "29bb79e96b047d79056626d747baa0b3a6ca9d4da7ad80b9668c0085e5581706" },
     { "id": "pilot-suite", "path": "benchmarks/production-v1/suite.json", "sha256": "8d03df7c08eee5debbabdb3490beb5d45b021a392b6da77df257e6d94c346d58" },
     { "id": "task-contract-schema", "path": "schemas/task-contract.schema.json", "sha256": "432a253af865f5c79527b3c7e5e65fc3b6cfea43d19f9e915f3bda792a358bb4" },
     { "id": "assurance-schema", "path": "schemas/benchmark-assurance-evidence.schema.json", "sha256": "5d937632d2cfff34dbbb781dd367bb78feb64c36d571c36ca15f2cc40c576af8" },

package/evals/private-holdout-v1/public-exposure.v1.json

@@ -3,7 +3,7 @@
   "id": "author-visible-evaluation-exposure-v1",
   "taxonomy": {
     "path": "evals/real-task-taxonomy.v1.json",
-    "sha256": "b3fa32c7008ce7b60de4ffebd6853da1601effae8b422f1cfa68810e4ea2150a"
+    "sha256": "d42458d04861a60488a5f98532936db71d43dcccc0298d75b09c2255f2922cb1"
   },
   "visibleSuites": [
     {
@@ -13,8 +13,8 @@
     },
     {
       "id": "capability-v1",
-      "manifestSha256": "378133ad8d10a47e580ecb62b1c0221d562d81a49c8992f3ab8caa46f5980ddc",
-      "scenarioIds": ["multi-package-rollout", "fullstack-search-contract", "concurrent-lease-lifecycle", "resumable-migration-runner"]
+      "manifestSha256": "29bb79e96b047d79056626d747baa0b3a6ca9d4da7ad80b9668c0085e5581706",
+      "scenarioIds": ["multi-package-rollout", "fullstack-search-contract", "concurrent-lease-lifecycle", "resumable-migration-runner", "durable-session-control-plane", "causal-timeline-recovery"]
     },
     {
       "id": "production-v1",

package/evals/real-task-taxonomy.v1.json

@@ -159,7 +159,11 @@
       "frameworks": ["node-workspace", "react", "nextjs"],
       "lifecycles": ["fresh", "repair"],
       "risks": ["routine", "high-risk"],
-      "historicalPublicScenarios": [{ "suite": "capability-v1", "id": "fullstack-search-contract", "evidenceTier": "historical-capability" }],
+      "historicalPublicScenarios": [
+        { "suite": "capability-v1", "id": "fullstack-search-contract", "evidenceTier": "historical-capability" },
+        { "suite": "capability-v1", "id": "durable-session-control-plane", "evidenceTier": "historical-capability" },
+        { "suite": "capability-v1", "id": "causal-timeline-recovery", "evidenceTier": "historical-capability" }
+      ],
       "nextEvidenceWorkItem": "CF-FS4-02"
     },
     {

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@piagent/platform",
-  "version": "1.5.1",
+  "version": "1.5.3",
   "license": "MIT",
   "publishConfig": {
     "access": "public",
@@ -157,6 +157,8 @@
     "init-project": "bash scripts/init-project.sh",
     "doctor": "bash scripts/team-doctor.sh",
     "benchmark": "node scripts/benchmark-runner.mjs",
+    "benchmark:deep": "node scripts/benchmark-runner.mjs --deep --surfaces piagent,codex-cli --model openai-codex/gpt-5.6-luna --thinking medium --allow-pi-auth-writeback",
+    "benchmark:webui-parity": "node packages/piagent-webui/benchmark/parity-benchmark.mjs",
     "benchmark:redaction": "node scripts/redaction-benchmark.mjs",
     "audit:runtime": "bash scripts/audit-runtime-host.sh",
     "release:identity": "node scripts/verify-release-identity.mjs",

package/packages/piagent-core/benchmark/benchmark-bootstrap.js

@@ -15,7 +15,7 @@ import {
 import { benchmarkTreeIdentity } from "./benchmark-tree-identity.js";
 import { benchmarkGitEnvironment, benchmarkHostEnvironment } from "./benchmark-runtime.js";
 
-const builtInSuites = new Set(["core-v1", "capability-v1", "e2-framework-v1", "production-v1"]);
+const builtInSuites = new Set(["core-v1", "capability-v1", "e2-framework-v1", "deep-logic-v1", "production-v1"]);
 const metadataVariable = "PIAGENT_BENCHMARK_BOOTSTRAP_METADATA";
 
 function fail(message) {
@@ -42,7 +42,7 @@ function jsonFile(file, label) {
   catch (error) { fail(`Cannot read ${label} ${file}: ${error.message}`); }
 }
 
-function requestedSuite(argv, cwd, replaySnapshot) {
+export function requestedSuite(argv, cwd, replaySnapshot) {
   const resume = optionValue(argv, "--resume");
   if (resume) {
     const manifestPath = path.join(runRootFromResume(resume, cwd), "run-manifest.json");
@@ -56,6 +56,7 @@ function requestedSuite(argv, cwd, replaySnapshot) {
     return replaySnapshot?.manifest?.suite?.source ?? report?.suite?.source ?? report?.suite?.manifestPath ?? report?.suite?.id ?? "production-v1";
   }
   if (argv.includes("--production")) return "production-v1";
+  if (argv.includes("--deep")) return "deep-logic-v1";
   if (argv.includes("--capability")) return "capability-v1";
   return optionValue(argv, "--suite") ?? "core-v1";
 }

package/packages/piagent-core/benchmark/benchmark-cli.js

@@ -16,7 +16,8 @@ With no options it uses the built-in core-v1 suite, the current Pi default
 model/thinking setting, and the suite's repeat count.
 
 Options:
-  --suite <id|path>            core-v1, capability-v1, e2-framework-v1, production-v1, or suite.json path.
+  --suite <id|path>            core-v1, capability-v1, e2-framework-v1, deep-logic-v1, production-v1, or suite.json path.
+  --deep                       Alias for --suite deep-logic-v1 (6 large scenarios, current model by default).
   --capability                 Alias for --suite capability-v1.
   --production                 Alias for --suite production-v1.
   --surfaces <a,b>             raw-pi,piagent (default) or piagent,codex-cli.
@@ -110,6 +111,9 @@ export function parseBenchmarkArgs(argv) {
       case "--production":
         options.suite = "production-v1";
         break;
+      case "--deep":
+        options.suite = "deep-logic-v1";
+        break;
       case "--capability":
         options.suite = "capability-v1";
         break;

package/packages/piagent-core/benchmark/benchmark-matrix.js

@@ -1,7 +1,7 @@
 export const BENCHMARK_SCOPE_BANDS = [
   { id: "core", suite: "core-v1", scenarios: 4, availability: "runnable", purpose: "Fast regression check for task quality, scope, safety, and exact usage." },
   { id: "production", suite: "production-v1", scenarios: 18, availability: "runnable", claimTier: "public-regression", purpose: "Broader public regression matrix across archetypes, lifecycle states, and verifier profiles; not a generalization or production-stability claim." },
-  { id: "capability", suite: "capability-v1", scenarios: 4, availability: "runnable", claimTier: "capability", purpose: "Unsaturated multi-file and multi-component tasks for hill-climbing; not a release gate." },
+  { id: "capability", suite: "capability-v1", scenarios: 6, availability: "runnable", claimTier: "capability", purpose: "Unsaturated multi-file and multi-component tasks for hill-climbing; not a release gate." },
   { id: "e2-framework", suite: "e2-framework-v1", scenarios: 4, availability: "runnable", claimTier: "capability", purpose: "Taxonomy-bound real Hono, SQLite, and workspace tasks with reference, mutation, alternative-valid, scope, and grader-sensitivity calibration; not a release gate." },
   { id: "long-horizon", suite: undefined, scenarios: 1, availability: "runnable-provider-free", evidenceLane: "evals/long-horizon-v1/lane.json", purpose: "At-least-30-minute provider-free lifecycle lane with hard process death, resume, compaction, handoff, bounded continuation, stable verification, and state-growth telemetry; not model-performance evidence." },
   { id: "private-holdout", suite: undefined, scenarios: 6, availability: "external-custody-required", claimTier: "private-holdout", readinessMatrix: "evals/fs4-readiness-matrix.v1.json", purpose: "Family- and repository-disjoint E3 readiness boundary; actual prompts, graders, repositories, human scores, and custody evidence remain external and cannot be self-attested." }

package/packages/piagent-core/benchmark/benchmark-suite-runtime.js

@@ -22,6 +22,7 @@ export function loadBenchmarkSuite(input, packageRoot) {
     ["core-v1", path.join(packageRoot, "benchmarks", "core-v1", "suite.json")],
     ["capability-v1", path.join(packageRoot, "benchmarks", "capability-v1", "suite.json")],
     ["e2-framework-v1", path.join(packageRoot, "benchmarks", "e2-framework-v1", "suite.json")],
+    ["deep-logic-v1", path.join(packageRoot, "benchmarks", "deep-logic-v1", "suite.json")],
     ["production-v1", path.join(packageRoot, "benchmarks", "production-v1", "suite.json")]
   ]);
   const candidate = builtIn.get(input) ?? path.resolve(input);

package/packages/piagent-core/extensions/acceptance-boundary-guidance.js

@@ -0,0 +1,64 @@
+function normalizedText(value) {
+  return String(value ?? "")
+    .normalize("NFD")
+    .replace(/[\u0300-\u036f]/g, "")
+    .toLowerCase();
+}
+
+function uniqueStrings(values) {
+  return [...new Set(values.filter(Boolean))];
+}
+
+/**
+ * Derive bounded, auditable proof obligations for stateful boundary contracts.
+ *
+ * The signals come only from the operator request. They intentionally describe
+ * reusable test partitions instead of benchmark identities or expected code.
+ */
+export function acceptanceBoundaryProofGuidance(raw) {
+  const value = normalizedText(raw);
+  const guidance = [];
+  const replayContract = /\b(?:replay(?:ed|s)?|idempot(?:ent|ency))\b/.test(value);
+
+  if (
+    /\b(?:safe[-\s]+integer|number\.max_safe_integer|number\.issafeinteger)\b/.test(value)
+    && /\b(?:advance|bump|cursor|decrement|increment|next|revision|successor)\b|(?:\+|-)\s*1\b/.test(value)
+  ) {
+    guidance.push("For safe-integer state transitions, exercise Number.MAX_SAFE_INTEGER and validate the computed next value before storing it; validating only the pre-transition input is insufficient.");
+  }
+
+  if (/\bfirst[-\s]+(?:observed|seen|encountered)[-\s]+order\b|\border\s+(?:first\s+)?(?:observed|seen|encountered)\b/.test(value)) {
+    guidance.push("Prove first-observed ordering with interleaved identities A, B, B, A; deduplication or replay detection must not reorder output by later duplicate positions.");
+  }
+
+  if (
+    replayContract
+    && /\b(?:conflict|different\s+content|different\s+command|identity|reuse|same\s+id|same\s+key|same-key)\b/.test(value)
+  ) {
+    guidance.push("Exercise fresh input, exact replay, and reuse of the same identity with different content as separate partitions; replay acceptance must not hide an identity conflict.");
+  }
+
+  if (
+    replayContract
+    && /\b(?:before\s+(?:the\s+)?revision|prior\s+idempotency\s+receipt|stale\s+(?:expected\s+)?revision)\b/.test(value)
+  ) {
+    guidance.push("When replay precedes revision matching, use a stale revision and prove the revision check does not reject an otherwise exact replay.");
+  }
+
+  if (replayContract && /\b(?:identical|same|exact)[-\s]+state[-\s]+object\b/.test(value)) {
+    guidance.push("Where replay promises state identity, assert the returned state is the exact prior object and that its stored revision remains unchanged.");
+  }
+
+  if (/\b(?:non[- ]whitespace|whitespace[- ]only|trim(?:med|ming)?)\b/.test(value)) {
+    guidance.push("Treat empty and whitespace-only strings as separate invalid partitions whenever the contract trims or requires a non-whitespace character.");
+  }
+
+  if (
+    /\b(?:backpressure|budget|capacity|maxchars)\b/.test(value)
+    && /\b(?:atomic(?:ally)?|equal(?:ity)?|exceed|remainder|unchanged)\b/.test(value)
+  ) {
+    guidance.push("Exercise capacity one below, at exact equality, and one above; the rejected item must leave coupled state unchanged and remain buffered when required.");
+  }
+
+  return uniqueStrings(guidance).slice(0, 6);
+}

package/packages/piagent-core/extensions/acceptance-contract-semantics.js

@@ -1,4 +1,5 @@
 import path from "node:path";
+import { acceptanceBoundaryProofGuidance } from "./acceptance-boundary-guidance.js";
 import { evidenceTopLevelArguments, executableRejectionAssertions } from "./acceptance-executable-evidence.js";
 
 const INTEGER_TARGET_STOPWORDS = new Set([
@@ -947,7 +948,7 @@ export function acceptanceInvalidInputEvidence(input = {}) {
 
 export function acceptanceContractProofGuidance(raw) {
   const value = normalizedText(raw);
-  const guidance = [];
+  const guidance = [...acceptanceBoundaryProofGuidance(raw)];
   if (onlyUndefinedContract(raw)) guidance.push("Prove undefined falls through while null, false, 0, and empty string are each preserved at the highest-precedence position.");
   if (/\btypeerror\b/.test(value)) guidance.push("Assert TypeError for every rejected partition named by the request; a different error class is not equivalent.");
   const integerTargets = integerConstraintTargets(raw);
@@ -958,7 +959,7 @@ export function acceptanceContractProofGuidance(raw) {
   if (/\bnon-negative\b/.test(value)) guidance.push("Exercise zero as valid and a negative value as invalid for every non-negative constraint.");
   if (/\binclusive\b|\bthrough\b|\bfrom\b[^.\n]{0,80}\bto\b/.test(value)) guidance.push("Exercise both inclusive endpoints and the nearest value outside each endpoint.");
   if (/\bround(?:ing)?\b|\bceil(?:ing)?\b|\bclamp\b/.test(value)) guidance.push("Exercise zero, exact, partial/rounding, below-minimum, and above-maximum behavior where applicable.");
-  return uniqueStrings(guidance).slice(0, 6);
+  return uniqueStrings(guidance).slice(0, 8);
 }
 
 export function acceptanceContractSemanticConflicts(obligation, taskText, sourceText) {

package/packages/piagent-core/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@piagent/core",
-  "version": "1.5.1",
+  "version": "1.5.3",
   "type": "module",
   "private": true,
   "license": "MIT",