Audit joint outcome network integrity

What it's for

Turns the site's implemented engineering-outcomes Bayesian network into a governable company-calibrated capability rather than an unexplained fixed prior.

What you give it

Inputs split into evidence read from your connected systems, calibration your team owns, and numerical controls that affect precision but never the result's meaning.

{
  "as_of_ms": 150,
  "conditional_probability_rows": [
    {
      "effective_sample_size": 70,
      "estimated_through_ms": 90,
      "evidence_verified": true,
      "id": "cpt-load",
      "network_version_id": "joint-network-v1",
      "node_id": "load",
      "parent_states": {},
      "probabilities": {
        "bad": 0.4,
        "good": 0.6
      },
      "source_artifact_id": "joint-training",
      "support_count": 100
    },
    {
      "effective_sample_size": 35,
      "estimated_through_ms": 90,
      "evidence_verified": true,
      "id": "cpt-delivery-good",
      "network_version_id": "joint-network-v1",
      "node_id": "delivery",
      "parent_states": {
        "load": "good"
      },
      "probabilities": {
        "bad": 0.19999999999999996,
        "good": 0.8
      },
      "source_artifact_id": "joint-training",
      "support_count": 50
    },
    {
      "effective_sample_size": 35,
      "estimated_through_ms": 90,
      "evidence_verified": true,
      "id": "cpt-delivery-bad",
      "network_version_id": "joint-network-v1",
      "node_id": "delivery",
      "parent_states": {
        "load": "bad"

What you get back

This is the actual output of running the example above — computed by the same function the platform calls, not an illustration.

{
  "configuration": {
    "as_of_ms": 150,
    "maximum_expected_calibration_error": 0.1,
    "minimum_effective_support": 20,
    "minimum_log_loss_improvement": 0,
    "minimum_outcome_maturity": 0.95,
    "minimum_validation_rows": 50
  },
  "decision": "pass",
  "failed_gates": {
    "network_failure_counts": {},
    "orphan_cpt_ids": [],
    "orphan_edge_ids": [],
    "orphan_node_ids": [],
    "orphan_validation_ids": []
  },
  "guardrails": [
    "Passing verifies the submitted DAG, CPT, temporal lineage, support and out-of-time validation; it does not prove that predictive edges are causal or that the model is complete.",
    "Protected traits, inferred intent and named-person employment scoring are outside this aggregate outcome-network contract.",
    "Passing does not authorize an employment, investment, procurement, rollout, access or production decision."
  ],
  "method": "point_in_time_joint_outcome_network_integrity_audit_v1",
  "network_diagnostics": [
    {
      "cpt_row_count": 5,
      "edge_count": 2,
      "failed_gates": [],
      "network_version_id": "joint-network-v1",
      "node_count": 3,
      "outcome_count": 2,
      "passed": true,
      "passing_validation_outcome_count": 2,
      "scope_id": "company"
    }
  ],
  "summary": {
    "conditional_probability_row_count": 5,
    "edge_count": 2,
    "network_version_count": 1,
    "node_count": 3,
    "passing_network_version_count": 1,
    "validation_slice_count": 2
  }

How it works

Sequential Bayesian & bandits — Learn while deciding — update beliefs as evidence arrives and choose where the next unit of effort is worth spending.

1. 1 Freeze the effective network version, intended use, owners, outcome nodes, training cutoff, artifacts and prohibited attributes at the requested as-of time.
2. 2 Validate the acyclic graph and every parent-state CPT row, including exact state coverage, normalization, temporal lineage, independent evidence and effective sample support.
3. 3 Require every declared outcome to pass a strictly later mature validation slice on sample size, log-loss improvement and expected calibration error.

Before you trust it

Every tool in the catalog ships with the conditions under which its answer is meaningful — and the conditions under which it should abstain instead of guessing.

Calibrating it to your org

Calibration workflow

1. 1 Define the management decision, target outcome, aggregate unit, privacy boundary, cadence, and prediction/intervention horizon for this organization.
2. 2 Build a tenant-scoped historical cohort using only information available before each prediction or decision; preserve zero periods, censoring, assignment probabilities, and unresolved outcomes when the method requires them.
3. 3 Estimate statistical parameters on training history, but obtain costs, utilities, risk tolerance, practical-effect thresholds, capacity, and policy constraints from accountable decision owners.
4. 4 Validate on later time windows or held-out aggregate units at the deployment grain, against a simple baseline and the function-specific validation strategy.
5. 5 Deploy only if the returned decision clears evidence, overlap, calibration, robustness, and guardrail checks; warning, unsupported, schema-gap, and fallback decisions are abstentions.
6. 6 Monitor realized outcomes, data drift, coverage, and decision regret; recalibrate at a governed cadence or after a detected regime/definition change, never merely because a stakeholder dislikes the result.

Call it from your AI

You don't wire up 388 tools in your MCP client. The GitRevio MCP server exposes 18 tools, three of which let an agent search the catalog, read a tool's schema, and run it — so the assistant finds this one on its own.

gitrevio_capabilities_search
  { "q": "audit whether a companyspecific bayesian jointoutcome" }
  → finds "audit_joint_outcome_network_integrity"

gitrevio_capability_describe
  { "capability_id": "audit_joint_outcome_network_integrity" }
  → returns the input schema and agent guidance shown on this page

gitrevio_capability_run
  { "capability_id": "audit_joint_outcome_network_integrity", "arguments": { ... } }
  → returns the result shown above

Works in Claude Desktop, Claude Code, Cursor, Cline, Continue.dev, Goose and Aider. See the MCP server.

Related tools

See every tool in Causal evidence & experiments →