Bayesian account risk triage

What it's for

Ranks account-security and policy-conflict evidence for human review by Bayes factor and the cost of being wrong, so the queue reflects risk rather than alert volume.

Adds defensible insider-risk and conflict-of-commitment triage without identity or nationality inference.

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.

{
  "cases": [
    {
      "case_id": "account-17",
      "impact_if_missed": 500000,
      "investigation_cost": 2000,
      "observations": [
        {
          "evidence_id": "iam-alert-1",
          "false_positive_rate": 0.01,
          "independence_group": "identity_access",
          "observed": true,
          "sensitivity": 0.75,
          "signal": "impossible_travel"
        }
      ],
      "prior_probability": 0.01
    }
  ]
}

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.

{
  "governance": [
    "This is an account-risk triage model, not an identity, nationality, espionage, or guilt classifier.",
    "Only auditable security events and explicit policy conflicts may be used; protected traits and nationality proxies are prohibited.",
    "Investigation requires human review, corroboration, access controls, due process, and a recorded disposition; the model takes no automatic adverse action.",
    "Concurrent employment is relevant only when an explicit policy applies and verified commitments conflict; asynchronous work patterns alone are not evidence."
  ],
  "method": "cost_sensitive_bayesian_risk_triage_v1",
  "triage_queue": [
    {
      "case_id": "account-17",
      "decision_threshold": 0.004,
      "evidence_groups": [
        {
          "capped_log_likelihood_ratio": 4.3175,
          "correlation_cap_applied": false,
          "evidence_ids": [
            "iam-alert-1"
          ],
          "independence_group": "identity_access",
          "raw_log_likelihood_ratio": 4.3175
        }
      ],
      "evidence_ids": [
        "iam-alert-1"
      ],
      "expected_avoided_loss": 215517.24,
      "investigation_cost": 2000,
      "net_value_of_investigation": 213517.24,
      "posterior_probability": 0.431,
      "prior_probability": 0.01,
      "recommendation": "human_investigation"
    }
  ]
}

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 Prioritize human review of auditable account-security and policy-conflict evidence using Bayes factors and decision costs.
2. 2 Evaluate the method-specific diagnostics and gates returned by the function, then abstain unless the declared decision clears them under locally governed thresholds.

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": "prioritize human review of auditable accountsecurity" }
  → finds "bayesian_account_risk_triage"

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

gitrevio_capability_run
  { "capability_id": "bayesian_account_risk_triage", "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 Security, access & compliance →