{
  "schemaVersion": "cg-agent-intent-conformance-result-0.1.0",
  "benchmarkId": "CG-BENCHMARK-AGENT-INTENT-CONFORMANCE-0.1.0",
  "version": "0.1.0",
  "testedAt": "2026-07-30T00:00:00.000Z",
  "engine": {
    "name": "CapitalGuard Agent Intent Firewall",
    "version": "0.1.0"
  },
  "methodologyVersion": "0.1.0",
  "fixtureDigestSha256": "8e2683cfa9d4e490ffd0348ce4be1df0d708ecdb6ad0c3f65abea931120a60e1",
  "caseCount": 15,
  "passedCaseCount": 15,
  "failedCaseCount": 0,
  "summary": {
    "acceptedCases": 1,
    "rejectedCases": 11,
    "exactlyOneAcceptedCases": 2,
    "uncertainConsumedCases": 1,
    "totalAttempts": 19,
    "downstreamInvocations": 5
  },
  "cases": [
    {
      "id": "CGAIC-TV-001",
      "title": "Approved read then write sequence",
      "boundary": "declared plan",
      "mutation": "none",
      "expected": "accepted",
      "signal": "ordered-plan-completed",
      "observed": "accepted",
      "attempts": 2,
      "acceptedCount": 2,
      "rejectedCount": 0,
      "downstreamInvocations": 2,
      "expectedCode": null,
      "observedCode": null,
      "passed": true
    },
    {
      "id": "CGAIC-TV-002",
      "title": "Signature threshold is incomplete",
      "boundary": "approval threshold",
      "mutation": "one of two required signatures omitted",
      "expected": "rejected",
      "signal": "contract-threshold-not-met",
      "observed": "rejected",
      "attempts": 1,
      "acceptedCount": 0,
      "rejectedCount": 1,
      "downstreamInvocations": 0,
      "expectedCode": "contract_threshold_not_met",
      "observedCode": "contract_threshold_not_met",
      "passed": true
    },
    {
      "id": "CGAIC-TV-003",
      "title": "Contract contains an untrusted signature",
      "boundary": "trusted approvers",
      "mutation": "second signature created by an unknown authority",
      "expected": "rejected",
      "signal": "untrusted-signature-blocked",
      "observed": "rejected",
      "attempts": 1,
      "acceptedCount": 0,
      "rejectedCount": 1,
      "downstreamInvocations": 0,
      "expectedCode": "untrusted_contract_signature",
      "observedCode": "untrusted_contract_signature",
      "passed": true
    },
    {
      "id": "CGAIC-TV-004",
      "title": "Signed objective digest is changed",
      "boundary": "contract integrity",
      "mutation": "objective digest changed after signing",
      "expected": "rejected",
      "signal": "contract-tamper-blocked",
      "observed": "rejected",
      "attempts": 1,
      "acceptedCount": 0,
      "rejectedCount": 1,
      "downstreamInvocations": 0,
      "expectedCode": "contract_digest_invalid",
      "observedCode": "contract_digest_invalid",
      "passed": true
    },
    {
      "id": "CGAIC-TV-005",
      "title": "Request actor differs from signed actor",
      "boundary": "actor binding",
      "mutation": "request actor changed",
      "expected": "rejected",
      "signal": "actor-mismatch-blocked",
      "observed": "rejected",
      "attempts": 1,
      "acceptedCount": 0,
      "rejectedCount": 1,
      "downstreamInvocations": 0,
      "expectedCode": "request_binding_mismatch",
      "observedCode": "request_binding_mismatch",
      "passed": true
    },
    {
      "id": "CGAIC-TV-006",
      "title": "Contract is used after expiry",
      "boundary": "time binding",
      "mutation": "verification time moved to contract expiry",
      "expected": "rejected",
      "signal": "expired-contract-blocked",
      "observed": "rejected",
      "attempts": 1,
      "acceptedCount": 0,
      "rejectedCount": 1,
      "downstreamInvocations": 0,
      "expectedCode": "contract_expired",
      "observedCode": "contract_expired",
      "passed": true
    },
    {
      "id": "CGAIC-TV-007",
      "title": "Request names an undeclared step",
      "boundary": "closed graph",
      "mutation": "step identifier absent from signed graph",
      "expected": "rejected",
      "signal": "unknown-step-blocked",
      "observed": "rejected",
      "attempts": 1,
      "acceptedCount": 0,
      "rejectedCount": 1,
      "downstreamInvocations": 0,
      "expectedCode": "step_not_found",
      "observedCode": "step_not_found",
      "passed": true
    },
    {
      "id": "CGAIC-TV-008",
      "title": "Write runs before its read dependency",
      "boundary": "dependency order",
      "mutation": "dependent step invoked first",
      "expected": "rejected",
      "signal": "out-of-order-step-blocked",
      "observed": "rejected",
      "attempts": 1,
      "acceptedCount": 0,
      "rejectedCount": 1,
      "downstreamInvocations": 0,
      "expectedCode": "dependency_incomplete",
      "observedCode": "dependency_incomplete",
      "passed": true
    },
    {
      "id": "CGAIC-TV-009",
      "title": "Write path differs from approved path",
      "boundary": "file target",
      "mutation": "approved output path replaced",
      "expected": "rejected",
      "signal": "path-substitution-blocked",
      "observed": "rejected",
      "attempts": 1,
      "acceptedCount": 0,
      "rejectedCount": 1,
      "downstreamInvocations": 0,
      "expectedCode": "resource_not_authorized",
      "observedCode": "resource_not_authorized",
      "passed": true
    },
    {
      "id": "CGAIC-TV-010",
      "title": "Write content digest differs",
      "boundary": "content binding",
      "mutation": "unreviewed content digest supplied",
      "expected": "rejected",
      "signal": "content-substitution-blocked",
      "observed": "rejected",
      "attempts": 1,
      "acceptedCount": 0,
      "rejectedCount": 1,
      "downstreamInvocations": 0,
      "expectedCode": "resource_not_authorized",
      "observedCode": "resource_not_authorized",
      "passed": true
    },
    {
      "id": "CGAIC-TV-011",
      "title": "Command arguments differ",
      "boundary": "command target",
      "mutation": "approved arguments replaced with force push",
      "expected": "rejected",
      "signal": "argument-substitution-blocked",
      "observed": "rejected",
      "attempts": 1,
      "acceptedCount": 0,
      "rejectedCount": 1,
      "downstreamInvocations": 0,
      "expectedCode": "resource_not_authorized",
      "observedCode": "resource_not_authorized",
      "passed": true
    },
    {
      "id": "CGAIC-TV-012",
      "title": "Network destination differs",
      "boundary": "network target",
      "mutation": "approved host replaced",
      "expected": "rejected",
      "signal": "destination-substitution-blocked",
      "observed": "rejected",
      "attempts": 1,
      "acceptedCount": 0,
      "rejectedCount": 1,
      "downstreamInvocations": 0,
      "expectedCode": "resource_not_authorized",
      "observedCode": "resource_not_authorized",
      "passed": true
    },
    {
      "id": "CGAIC-TV-013",
      "title": "Identical request is replayed",
      "boundary": "request replay",
      "mutation": "same request submitted twice in sequence",
      "expected": "exactly_one_accepted",
      "signal": "sequential-replay-blocked",
      "observed": "exactly_one_accepted",
      "attempts": 2,
      "acceptedCount": 1,
      "rejectedCount": 1,
      "downstreamInvocations": 1,
      "expectedCode": "request_replay",
      "observedCode": "request_replay",
      "passed": true
    },
    {
      "id": "CGAIC-TV-014",
      "title": "Two requests race for one allowance",
      "boundary": "atomic allowance",
      "mutation": "two distinct requests submitted concurrently",
      "expected": "exactly_one_accepted",
      "signal": "concurrent-limit-enforced",
      "observed": "exactly_one_accepted",
      "attempts": 2,
      "acceptedCount": 1,
      "rejectedCount": 1,
      "downstreamInvocations": 1,
      "expectedCode": "step_limit_exceeded",
      "observedCode": "step_limit_exceeded",
      "passed": true
    },
    {
      "id": "CGAIC-TV-015",
      "title": "Timeout after dispatch consumes allowance",
      "boundary": "uncertain outcome",
      "mutation": "downstream throws after invocation begins, then retry attempted",
      "expected": "uncertain_consumed",
      "signal": "uncertain-outcome-not-replayed",
      "observed": "uncertain_consumed",
      "attempts": 2,
      "acceptedCount": 0,
      "rejectedCount": 2,
      "downstreamInvocations": 1,
      "expectedCode": "downstream_outcome_uncertain,step_limit_exceeded",
      "observedCode": "downstream_outcome_uncertain,step_limit_exceeded",
      "passed": true
    }
  ],
  "claimBoundary": "This publisher-run deterministic benchmark tests CapitalGuard Agent Intent Firewall 0.1.0 against fifteen disclosed synthetic plan-conformance cases. It demonstrates only the observed handling of threshold signatures, bindings, expiry, declared step order, exact selectors, replay, atomic allowance, and uncertain outcomes for operations routed through the owned callback. It does not infer whether prose matches a human objective, prevent every prompt injection or goal-hijack technique, observe bypassed operations, prove production deployment quality, or compare CapitalGuard with another product.",
  "limitations": [
    "The fixtures use synthetic identities, digests, keys, and inert callbacks; no model, external agent, customer repository, shell command, file mutation, or network request is executed.",
    "A signed graph can still encode a poor or malicious plan. The benchmark tests exact conformance to that graph, not whether reviewers understood or approved a good objective.",
    "Operations that bypass the owned callback are outside the measured boundary.",
    "The public verifier checks fixture, signature, and results consistency; it does not independently attest the publisher-run engine execution.",
    "Host, authority-key, state-authority, or coordinated state-and-anchor compromise is outside this local benchmark."
  ],
  "privacy": {
    "customerDataUsed": false,
    "productionKeysUsed": false,
    "networkActionsExecuted": false,
    "externalAgentsExecuted": false,
    "shellCommandsExecuted": false,
    "fileMutationsExecuted": false,
    "sourceContentsIncluded": false,
    "absolutePathsIncluded": false
  },
  "runDigestSha256": "6b72658b93e1e44a0567d72d38c878c2504d3f47d78364afb42f0bd6718c50fe"
}