@raishin/vanguard-frontier-agentic 3.15.0 → 3.17.0

package/tests/agentic-board-suite/fixtures/invalid/STOP-CHALLENGE-ROUNDS--third-version-by-return-naming-the-original.json

@@ -0,0 +1,461 @@
+{
+  "description": "Sole violation of STOP-CHALLENGE-ROUNDS: third version by return naming the original (derived from v11).",
+  "messages": [
+    {
+      "schema_version": "1.0.0",
+      "message_type": "delegate",
+      "task_id": "v11-m1",
+      "correlation_id": "v11-corr",
+      "parent_task_id": null,
+      "hop_count": 0,
+      "max_hops": 8,
+      "sent_at": "2026-10-01T09:01:00Z",
+      "source": {
+        "kind": "principal",
+        "subject": "subject:corp-idp:7c1e9a40b2d35f68",
+        "role": "platform-owner"
+      },
+      "destination": {
+        "kind": "agent",
+        "id": "agentic-system-architecture-review-agent"
+      },
+      "objective": "Assess whether the supplier-onboarding multi-agent design is fit for a non-production pilot.",
+      "input_schema_version": "1.0.0",
+      "subject_risk_tier": "R2",
+      "data_classification": "internal",
+      "materiality_flags": [],
+      "authorized_actions": [
+        "read",
+        "analyze",
+        "recommend",
+        "challenge",
+        "escalate"
+      ],
+      "do_not_do_list": [
+        "Do not execute, deploy, or mutate any system",
+        "Do not request, store, or echo credentials, tokens, or secrets",
+        "Do not record an approval on behalf of a human"
+      ],
+      "evidence_refs": [
+        {
+          "ref": "repo:docs/design/v11.md#section-1",
+          "label": "repo-evidence"
+        }
+      ],
+      "missing_evidence": [],
+      "budget": {
+        "currency": "USD",
+        "limit_tool_calls": 40,
+        "spent_tool_calls": 0,
+        "limit_tokens": 400000,
+        "spent_tokens": 0,
+        "limit_cost_minor": 2000,
+        "spent_cost_minor": 0
+      },
+      "deadline": "2026-10-08T17:00:00Z",
+      "approval_state": "not-required",
+      "expected_output_schema": "agentic-review-verdict/1.0.0",
+      "status": "dispatched",
+      "stop_reasons": [],
+      "synthetic": true,
+      "decision_class": "architecture",
+      "decision_owner": {
+        "kind": "principal",
+        "subject": "subject:corp-idp:a94f03d2c8e1b756",
+        "role": "ai-risk-owner"
+      }
+    },
+    {
+      "schema_version": "1.0.0",
+      "message_type": "delegate",
+      "task_id": "v11-m2",
+      "correlation_id": "v11-corr",
+      "parent_task_id": "v11-m1",
+      "hop_count": 1,
+      "max_hops": 8,
+      "sent_at": "2026-10-01T09:02:00Z",
+      "source": {
+        "kind": "agent",
+        "id": "agentic-system-architecture-review-agent"
+      },
+      "destination": {
+        "kind": "agent",
+        "id": "agentic-security-review-agent"
+      },
+      "objective": "Challenge the proposal for v11: the design forwards user tokens to the ERP MCP server.",
+      "input_schema_version": "1.0.0",
+      "subject_risk_tier": "R3",
+      "data_classification": "internal",
+      "materiality_flags": [
+        "security-critical"
+      ],
+      "authorized_actions": [
+        "read",
+        "analyze",
+        "recommend",
+        "challenge",
+        "escalate"
+      ],
+      "do_not_do_list": [
+        "Do not execute, deploy, or mutate any system",
+        "Do not request, store, or echo credentials, tokens, or secrets",
+        "Do not record an approval on behalf of a human"
+      ],
+      "evidence_refs": [
+        {
+          "ref": "repo:docs/design/v11.md#section-1",
+          "label": "repo-evidence"
+        }
+      ],
+      "missing_evidence": [],
+      "budget": {
+        "currency": "USD",
+        "limit_tool_calls": 40,
+        "spent_tool_calls": 8,
+        "limit_tokens": 400000,
+        "spent_tokens": 40000,
+        "limit_cost_minor": 2000,
+        "spent_cost_minor": 160
+      },
+      "deadline": "2026-10-08T17:00:00Z",
+      "approval_state": "not-required",
+      "expected_output_schema": "agentic-review-verdict/1.0.0",
+      "status": "dispatched",
+      "stop_reasons": [],
+      "synthetic": true
+    },
+    {
+      "schema_version": "1.0.0",
+      "message_type": "challenge",
+      "task_id": "v11-m3",
+      "correlation_id": "v11-corr",
+      "parent_task_id": "v11-m2",
+      "hop_count": 2,
+      "max_hops": 8,
+      "sent_at": "2026-10-01T09:03:00Z",
+      "source": {
+        "kind": "agent",
+        "id": "agentic-security-review-agent"
+      },
+      "destination": {
+        "kind": "agent",
+        "id": "agentic-system-architecture-review-agent"
+      },
+      "objective": "Challenge findings for the v11 proposal.",
+      "input_schema_version": "1.0.0",
+      "subject_risk_tier": "R3",
+      "data_classification": "internal",
+      "materiality_flags": [
+        "security-critical"
+      ],
+      "authorized_actions": [
+        "read",
+        "analyze",
+        "recommend",
+        "challenge",
+        "escalate"
+      ],
+      "do_not_do_list": [
+        "Do not execute, deploy, or mutate any system",
+        "Do not request, store, or echo credentials, tokens, or secrets",
+        "Do not record an approval on behalf of a human"
+      ],
+      "evidence_refs": [
+        {
+          "ref": "repo:docs/design/v11.md#section-1",
+          "label": "repo-evidence"
+        }
+      ],
+      "missing_evidence": [],
+      "budget": {
+        "currency": "USD",
+        "limit_tool_calls": 40,
+        "spent_tool_calls": 14,
+        "limit_tokens": 400000,
+        "spent_tokens": 70000,
+        "limit_cost_minor": 2000,
+        "spent_cost_minor": 280
+      },
+      "deadline": "2026-10-08T17:00:00Z",
+      "approval_state": "not-required",
+      "expected_output_schema": "agentic-review-verdict/1.0.0",
+      "status": "completed",
+      "stop_reasons": [],
+      "synthetic": true,
+      "challenge": {
+        "proposal_task_id": "v11-m2",
+        "round": 1,
+        "hard_gate": true,
+        "outcome": "sustained",
+        "finding": "Token passthrough: the ERP MCP server receives a token whose audience it cannot validate."
+      }
+    },
+    {
+      "schema_version": "1.0.0",
+      "message_type": "return",
+      "task_id": "v11-r1",
+      "correlation_id": "v11-corr",
+      "parent_task_id": "v11-m3",
+      "hop_count": 3,
+      "max_hops": 8,
+      "sent_at": "2026-10-01T09:04:00Z",
+      "source": {
+        "kind": "agent",
+        "id": "agentic-system-architecture-review-agent"
+      },
+      "destination": {
+        "kind": "agent",
+        "id": "agentic-security-review-agent"
+      },
+      "objective": "Return the revised v11 proposal to the challenger.",
+      "input_schema_version": "1.0.0",
+      "subject_risk_tier": "R3",
+      "data_classification": "internal",
+      "materiality_flags": [
+        "security-critical"
+      ],
+      "authorized_actions": [
+        "read",
+        "analyze",
+        "recommend",
+        "challenge",
+        "escalate"
+      ],
+      "do_not_do_list": [
+        "Do not execute, deploy, or mutate any system",
+        "Do not request, store, or echo credentials, tokens, or secrets",
+        "Do not record an approval on behalf of a human"
+      ],
+      "evidence_refs": [
+        {
+          "ref": "repo:docs/design/v11.md#section-1",
+          "label": "repo-evidence"
+        }
+      ],
+      "missing_evidence": [],
+      "budget": {
+        "currency": "USD",
+        "limit_tool_calls": 40,
+        "spent_tool_calls": 14,
+        "limit_tokens": 400000,
+        "spent_tokens": 70000,
+        "limit_cost_minor": 2000,
+        "spent_cost_minor": 280
+      },
+      "deadline": "2026-10-08T17:00:00Z",
+      "approval_state": "not-required",
+      "expected_output_schema": "agentic-review-verdict/1.0.0",
+      "status": "completed",
+      "stop_reasons": [],
+      "synthetic": true,
+      "verdict": {
+        "outcome": "pass-with-conditions",
+        "summary": "Revised proposal returned for another look.",
+        "conditions": [
+          "Keep the audience-bound token exchange"
+        ]
+      }
+    },
+    {
+      "schema_version": "1.0.0",
+      "message_type": "challenge",
+      "task_id": "v11-c2",
+      "correlation_id": "v11-corr",
+      "parent_task_id": "v11-r1",
+      "hop_count": 4,
+      "max_hops": 8,
+      "sent_at": "2026-10-01T09:05:00Z",
+      "source": {
+        "kind": "agent",
+        "id": "agentic-security-review-agent"
+      },
+      "destination": {
+        "kind": "agent",
+        "id": "agentic-system-architecture-review-agent"
+      },
+      "objective": "Challenge findings for the v11 proposal.",
+      "input_schema_version": "1.0.0",
+      "subject_risk_tier": "R3",
+      "data_classification": "internal",
+      "materiality_flags": [
+        "security-critical"
+      ],
+      "authorized_actions": [
+        "read",
+        "analyze",
+        "recommend",
+        "challenge",
+        "escalate"
+      ],
+      "do_not_do_list": [
+        "Do not execute, deploy, or mutate any system",
+        "Do not request, store, or echo credentials, tokens, or secrets",
+        "Do not record an approval on behalf of a human"
+      ],
+      "evidence_refs": [
+        {
+          "ref": "repo:docs/design/v11.md#section-1",
+          "label": "repo-evidence"
+        }
+      ],
+      "missing_evidence": [],
+      "budget": {
+        "currency": "USD",
+        "limit_tool_calls": 40,
+        "spent_tool_calls": 14,
+        "limit_tokens": 400000,
+        "spent_tokens": 70000,
+        "limit_cost_minor": 2000,
+        "spent_cost_minor": 280
+      },
+      "deadline": "2026-10-08T17:00:00Z",
+      "approval_state": "not-required",
+      "expected_output_schema": "agentic-review-verdict/1.0.0",
+      "status": "completed",
+      "stop_reasons": [],
+      "synthetic": true,
+      "challenge": {
+        "proposal_task_id": "v11-m2",
+        "round": 2,
+        "hard_gate": true,
+        "outcome": "sustained",
+        "finding": "Still token passthrough, answered against the original envelope."
+      }
+    },
+    {
+      "schema_version": "1.0.0",
+      "message_type": "return",
+      "task_id": "v11-r2",
+      "correlation_id": "v11-corr",
+      "parent_task_id": "v11-c2",
+      "hop_count": 5,
+      "max_hops": 8,
+      "sent_at": "2026-10-01T09:06:00Z",
+      "source": {
+        "kind": "agent",
+        "id": "agentic-system-architecture-review-agent"
+      },
+      "destination": {
+        "kind": "agent",
+        "id": "agentic-security-review-agent"
+      },
+      "objective": "Return the revised v11 proposal to the challenger.",
+      "input_schema_version": "1.0.0",
+      "subject_risk_tier": "R3",
+      "data_classification": "internal",
+      "materiality_flags": [
+        "security-critical"
+      ],
+      "authorized_actions": [
+        "read",
+        "analyze",
+        "recommend",
+        "challenge",
+        "escalate"
+      ],
+      "do_not_do_list": [
+        "Do not execute, deploy, or mutate any system",
+        "Do not request, store, or echo credentials, tokens, or secrets",
+        "Do not record an approval on behalf of a human"
+      ],
+      "evidence_refs": [
+        {
+          "ref": "repo:docs/design/v11.md#section-1",
+          "label": "repo-evidence"
+        }
+      ],
+      "missing_evidence": [],
+      "budget": {
+        "currency": "USD",
+        "limit_tool_calls": 40,
+        "spent_tool_calls": 14,
+        "limit_tokens": 400000,
+        "spent_tokens": 70000,
+        "limit_cost_minor": 2000,
+        "spent_cost_minor": 280
+      },
+      "deadline": "2026-10-08T17:00:00Z",
+      "approval_state": "not-required",
+      "expected_output_schema": "agentic-review-verdict/1.0.0",
+      "status": "completed",
+      "stop_reasons": [],
+      "synthetic": true,
+      "verdict": {
+        "outcome": "pass-with-conditions",
+        "summary": "Revised proposal returned for another look.",
+        "conditions": [
+          "Keep the audience-bound token exchange"
+        ]
+      }
+    },
+    {
+      "schema_version": "1.0.0",
+      "message_type": "challenge",
+      "task_id": "v11-c3",
+      "correlation_id": "v11-corr",
+      "parent_task_id": "v11-r2",
+      "hop_count": 6,
+      "max_hops": 8,
+      "sent_at": "2026-10-01T09:07:00Z",
+      "source": {
+        "kind": "agent",
+        "id": "agentic-security-review-agent"
+      },
+      "destination": {
+        "kind": "agent",
+        "id": "agentic-system-architecture-review-agent"
+      },
+      "objective": "Challenge findings for the v11 proposal.",
+      "input_schema_version": "1.0.0",
+      "subject_risk_tier": "R3",
+      "data_classification": "internal",
+      "materiality_flags": [
+        "security-critical"
+      ],
+      "authorized_actions": [
+        "read",
+        "analyze",
+        "recommend",
+        "challenge",
+        "escalate"
+      ],
+      "do_not_do_list": [
+        "Do not execute, deploy, or mutate any system",
+        "Do not request, store, or echo credentials, tokens, or secrets",
+        "Do not record an approval on behalf of a human"
+      ],
+      "evidence_refs": [
+        {
+          "ref": "repo:docs/design/v11.md#section-1",
+          "label": "repo-evidence"
+        }
+      ],
+      "missing_evidence": [],
+      "budget": {
+        "currency": "USD",
+        "limit_tool_calls": 40,
+        "spent_tool_calls": 14,
+        "limit_tokens": 400000,
+        "spent_tokens": 70000,
+        "limit_cost_minor": 2000,
+        "spent_cost_minor": 280
+      },
+      "deadline": "2026-10-08T17:00:00Z",
+      "approval_state": "not-required",
+      "expected_output_schema": "agentic-review-verdict/1.0.0",
+      "status": "completed",
+      "stop_reasons": [],
+      "synthetic": true,
+      "challenge": {
+        "proposal_task_id": "v11-m2",
+        "round": 3,
+        "hard_gate": true,
+        "outcome": "sustained",
+        "finding": "Still token passthrough, answered against the original envelope."
+      }
+    }
+  ],
+  "expected_violations": [
+    "STOP-CHALLENGE-ROUNDS"
+  ]
+}

package/tests/agentic-board-suite/fixtures/invalid/STOP-DEADLINE--sent-after-deadline.json

@@ -0,0 +1,132 @@
+{
+  "description": "Sole violation of STOP-DEADLINE: sent after deadline (derived from v01).",
+  "messages": [
+    {
+      "schema_version": "1.0.0",
+      "message_type": "delegate",
+      "task_id": "v01-m1",
+      "correlation_id": "v01-corr",
+      "parent_task_id": null,
+      "hop_count": 0,
+      "max_hops": 8,
+      "sent_at": "2026-10-01T09:01:00Z",
+      "source": {
+        "kind": "principal",
+        "subject": "subject:corp-idp:7c1e9a40b2d35f68",
+        "role": "platform-owner"
+      },
+      "destination": {
+        "kind": "agent",
+        "id": "agentic-system-architecture-review-agent"
+      },
+      "objective": "Assess whether a single tool-using agent or a deterministic workflow fits a read-only knowledge-base Q&A assistant.",
+      "input_schema_version": "1.0.0",
+      "subject_risk_tier": "R0",
+      "data_classification": "internal",
+      "materiality_flags": [],
+      "authorized_actions": [
+        "read",
+        "analyze",
+        "recommend",
+        "challenge",
+        "escalate"
+      ],
+      "do_not_do_list": [
+        "Do not execute, deploy, or mutate any system",
+        "Do not request, store, or echo credentials, tokens, or secrets",
+        "Do not record an approval on behalf of a human"
+      ],
+      "evidence_refs": [
+        {
+          "ref": "repo:docs/design/v01.md#section-1",
+          "label": "repo-evidence"
+        }
+      ],
+      "missing_evidence": [],
+      "budget": {
+        "currency": "USD",
+        "limit_tool_calls": 40,
+        "spent_tool_calls": 0,
+        "limit_tokens": 400000,
+        "spent_tokens": 0,
+        "limit_cost_minor": 2000,
+        "spent_cost_minor": 0
+      },
+      "deadline": "2026-10-08T17:00:00Z",
+      "approval_state": "not-required",
+      "expected_output_schema": "agentic-review-verdict/1.0.0",
+      "status": "dispatched",
+      "stop_reasons": [],
+      "synthetic": true
+    },
+    {
+      "schema_version": "1.0.0",
+      "message_type": "return",
+      "task_id": "v01-m2",
+      "correlation_id": "v01-corr",
+      "parent_task_id": "v01-m1",
+      "hop_count": 1,
+      "max_hops": 8,
+      "sent_at": "2026-10-08T17:00:01Z",
+      "source": {
+        "kind": "agent",
+        "id": "agentic-system-architecture-review-agent"
+      },
+      "destination": {
+        "kind": "principal",
+        "subject": "subject:corp-idp:7c1e9a40b2d35f68",
+        "role": "platform-owner"
+      },
+      "objective": "Return the pattern-selection verdict for the knowledge-base assistant review.",
+      "input_schema_version": "1.0.0",
+      "subject_risk_tier": "R0",
+      "data_classification": "internal",
+      "materiality_flags": [],
+      "authorized_actions": [
+        "read",
+        "analyze",
+        "recommend",
+        "challenge",
+        "escalate"
+      ],
+      "do_not_do_list": [
+        "Do not execute, deploy, or mutate any system",
+        "Do not request, store, or echo credentials, tokens, or secrets",
+        "Do not record an approval on behalf of a human"
+      ],
+      "evidence_refs": [
+        {
+          "ref": "repo:docs/design/v01.md#section-1",
+          "label": "repo-evidence"
+        }
+      ],
+      "missing_evidence": [],
+      "budget": {
+        "currency": "USD",
+        "limit_tool_calls": 40,
+        "spent_tool_calls": 6,
+        "limit_tokens": 400000,
+        "spent_tokens": 30000,
+        "limit_cost_minor": 2000,
+        "spent_cost_minor": 120
+      },
+      "deadline": "2026-10-08T17:00:00Z",
+      "approval_state": "not-required",
+      "expected_output_schema": "agentic-review-verdict/1.0.0",
+      "status": "completed",
+      "stop_reasons": [],
+      "synthetic": true,
+      "verdict": {
+        "outcome": "simpler-pattern-sufficient",
+        "summary": "A single retrieval agent with a fixed tool list meets the requirement; multi-agent adds handoff cost without separation-of-duties benefit.",
+        "conditions": [
+          "Bound tool calls per request",
+          "Log retrieval sources per answer"
+        ]
+      }
+    }
+  ],
+  "expected_violations": [
+    "STOP-DEADLINE"
+  ]
+}