Optimize CI assurance portfolio

What it's for

Answers the executive question CI dashboards avoid: which caching, sharding, flaky-test repair, mutation, integration-test or runner investment has causal evidence and creates the best risk-adjusted assurance within real money and capacity limits?

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.

{
  "assurance_units": [
    {
      "baseline_compute_minutes_per_run": 2,
      "baseline_detection_probability": 0.5,
      "baseline_feedback_minutes": 10,
      "common_loss_group_id": "checkout",
      "compute_cost_per_minute": 0.1,
      "evidence_verified": true,
      "expected_fault_count": 1,
      "expected_run_count": 40,
      "id": "checkout-ci",
      "pipeline_class_id": "fast",
      "required_control_ids": [
        "required-tests"
      ],
      "shared_common_loss": 20000,
      "undetected_fault_loss": 10000,
      "value_per_feedback_minute": 2
    }
  ],
  "budget": 5000,
  "common_scenarios": [
    {
      "arrival_multiplier": 1,
      "common_tool_failure_probability": 0,
      "fault_multiplier": 1,
      "id": "normal",
      "probability": 1,
      "runner_availability": 1,
      "service_time_multiplier": 1,
      "shared_loss_multiplier": 1
    }
  ],
  "implementation_capacity_hours": 40,
  "intervention_studies": [
    {
      "control_detected_count": 15,
      "control_fault_count": 30,
      "control_feedback_count": 30,
      "control_feedback_log_sq_sum": 129.72231375791435,
      "control_feedback_log_sum": 62.383246250395075,
      "evidence_verified": true,
      "id": "flake-switchback",

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.

{
  "constraints": {
    "budget": 5000,
    "implementation_capacity_hours": 40,
    "maximum_cvar_loss": 100000,
    "maximum_expected_undetected_faults": 5,
    "maximum_p95_feedback_minutes": 100,
    "maximum_runner_overload_probability": 0.05,
    "minimum_effective_sample_size": 20,
    "runner_capacity_minutes": 500,
    "tail_probability": 0.1
  },
  "decision": "selected",
  "frontier_truncated": false,
  "interpretation": "Selection is a human-governed planning recommendation over submitted options. It does not authorize skipping required tests, changing branch protection, merging code, or judging any contributor.",
  "method": "transport_weighted_causal_ci_assurance_stochastic_portfolio",
  "option_evidence": [
    {
      "assurance_unit_id": "checkout-ci",
      "effective_sample_size": null,
      "evidence_eligible": true,
      "option_id": "current",
      "option_type": "baseline",
      "probability_detection_improves": null,
      "probability_feedback_improves": null
    },
    {
      "assurance_unit_id": "checkout-ci",
      "effective_sample_size": 30,
      "evidence_eligible": true,
      "option_id": "repair-flakes",
      "option_type": "flaky_repair",
      "probability_detection_improves": 1,
      "probability_feedback_improves": 1
    }
  ],
  "pareto_frontier": [
    {
      "conditional_value_at_risk": 35653.5238,
      "cost": 0,
      "expected_loss": 13436.8571,
      "expected_undetected_faults": 0.5483,
      "option_ids": [
        "current"

How it works

Causal inference & experiment design — Separate what a change caused from what merely moved alongside it.

1. 1 Admit non-baseline effects only from prospectively registered randomized or known-propensity studies with injected/governed fault-detection denominators, mature feedback log moments, verified treatment versions and approved class transport similarity; shrink transported log effects toward no effect.
2. 2 Enumerate or deterministically beam-search one option per assurance unit, preserve required controls and option relations, and replay every alternative under common fault arrivals, detection uncertainty, runner/service shocks and shared common-loss groups while counting each shared loss once.
3. 3 Select the feasible minimum expected loss-plus-cost portfolio under budget, implementation, declared and stochastic runner capacity, expected undetected-fault, p95 feedback and CVaR limits; expose effect eligibility, solver certainty and a cost-risk Pareto frontier.

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": "choose one baseline cache shard testselection" }
  → finds "optimize_ci_assurance_portfolio"

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

gitrevio_capability_run
  { "capability_id": "optimize_ci_assurance_portfolio", "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 Quality, incidents & reliability →