design-mitigation

Purpose

Design an intervention that prevents, detects, responds to, removes, or relaxes a failure mode.

Input contract

required: [failure_mode, causal_mechanism, resource_limits]
optional: [existing_controls, timeline, acceptance_criteria]
constraints: [residual risk and validation evidence must be explicit]

Procedure

1. Classify the failure and locate controllable causal points.
2. Generate prevention, detection, response, removal, and relaxation options.
3. Sequence selected actions with resources and validation tests.
4. Estimate residual risk and define escalation conditions.

If the mitigation has an explicit mechanism and measurable target, consider validate-mitigation-effect as the next tactic.

Output contract

produces: [mitigation_plan, validation_tests, residual_risk, success_criteria]
delta_fields: [decisions, findings, uncertainties]

Quality gates

• Plan contains at least 3 sequenced actions when a multi-step intervention is feasible.
• Each action has an owner-independent resource statement, validation test, and success criterion.
• Residual risk is not reported as zero without evidence.

Parameterization

Caller supplies failure taxonomy, causal graph, intervention classes, resource schema, timeline, and risk scale.

Failure and counterexamples

Reject vague actions, controls that do not touch the mechanism, or plans without residual-risk accounting.

Provenance map

• concept: stress-test/mitigation-design-sop
• concept: stress-test/re-scoring
• concept: convergence/removal-path
• intermediate: Pass3/design-mitigation
• intermediate: Pass3/design-removal-path