finance.credit.scenario_set.discovery.v1
Overview
Verify that the Forge runtime executes the canonical IFRS 9 Scenario Set Discovery Primitive Profile using the production Search execution pipeline.
After completing this example you should be able to:
- execute the canonical profile
- inspect runtime evidence
- verify generated artifacts
- validate deterministic replay
- understand the execution contract
This document verifies an executable Search capability rather than IFRS 9 methodology itself.
Primitive Profile
| Property | Value |
|---|---|
| Primitive | search@1 |
| Profile | finance.credit.scenario_set.discovery.v1 |
| Runtime | Compute |
| Replay | Supported |
| Artifacts | Supported |
| Deterministic | Supported |
| CPU | Supported |
| GPU | Not Required |
| Validator | validateFinanceCreditIFRS9SearchArgs |
Capability
finance.credit.scenario_set.discovery.v1 searches the IFRS 9 macroeconomic scenario space to discover combinations of scenario assumptions that produce fragile or highly sensitive credit portfolio behaviour.
Rather than estimating a single Expected Credit Loss value, the profile explores alternative scenario configurations and ranks them according to a deterministic discovery objective.
Typical search dimensions include macroeconomic stress, GDP shocks, unemployment, interest-rate movements, scenario weighting, management overlays, calibration error, and volatility.
Canonical Contract
Execution uses the canonical Forge Search execution contract.
Primitive
search
Version
1
Profile
finance.credit.scenario_set.discovery.v1Requests are validated before execution begins.
Required Inputs
The canonical validator requires:
- iterations
- base_scenario
- mutation_rules
- ranking
- ranking.objective
- ranking.top_n
Execution cannot begin unless every required field satisfies canonical validation.
The validator additionally requires that the ranking objective matches the profile-specific discovery objective.
Optional Inputs
Supported optional scenario dimensions include:
- macro_stress
- GDP shock
- unemployment shock
- interest-rate shock
- scenario weight
- PD multiplier
- LGD multiplier
- EAD multiplier
- forward adjustment
- model overlay
- management overlay
- calibration error
- volatility
Mutation rules may be defined for supported dimensions using deterministic minimum, maximum, and step values.
Only documented mutation axes are accepted.
Unsupported mutation dimensions are rejected during validation.
Canonical Smoke
The maintained Forge smoke verifies:
- successful validation
- successful execution
- scenario discovery
- deterministic seed handling
- ranked candidate generation
- replay metadata
- execution artifacts
The maintained Smoke Suite remains the canonical source for executable payloads.
Verification Expectations
A successful execution should demonstrate:
- canonical validation succeeds
- profile-specific ranking objective is accepted
- mutation rules are accepted
- candidate scenarios are generated
- ranked scenario sets are returned
- runtime evidence is produced
- replay metadata is available
- execution artifacts are generated when requested
Replay should be verified after the initial execution.
Runtime Evidence
Successful execution exposes runtime evidence including:
- execution identifier
- primitive profile
- execution metadata
- selected discovery objective
- candidate count
- ranked scenario sets
- score distribution
- mutation trace
- replay metadata
- artifact references
The evidence surface is independent of the execution interface used to submit the workload.
Replay
Replay should preserve deterministic discovery behaviour when executed using the same canonical execution contract and deterministic seed.
Replay confirms the stability of the discovered scenario ranking rather than infrastructure identity.
Replay verification is described in:
→ /verification/replay-determinism
Artifacts
Typical execution artifacts include:
- ranked scenario sets
- execution summary
- mutation trace
- score distribution
- replay metadata
Additional artifacts may become available as the profile evolves.
Applied Intelligence Modules
This Primitive Profile is reused across multiple Forge Intelligence Modules, including:
- Credit Intelligence
- Banking Intelligence
- Portfolio Intelligence
- Financial Risk Intelligence
- Capital Intelligence
The computational contract remains identical regardless of the consuming Intelligence Module.
Related Documentation
- /examples/search/
- /verification/
- /verification/smoke-suite
- /verification/replay-determinism
- /verification/artifact-inspection
Verification Checklist
| Verification Surface | Status |
|---|---|
| Primitive resolved | ✓ |
| Contract validated | ✓ |
| Execution completed | ✓ |
| Runtime inspected | ✓ |
| Artifacts inspected | ✓ |
| Replay verified | ✓ |
| Negative validation tested | ✓ |
Final Principle
finance.credit.scenario_set.discovery.v1 verifies a versioned Search Primitive Profile whose behaviour is independently observable through canonical validation, deterministic execution, runtime evidence, replay, and artifact inspection.
The profile verifies the reproducibility of scenario discovery rather than the correctness of any individual macroeconomic assumption.
Continue in Forge Studio
This document describes the canonical execution contract for this capability.
The Forge documentation explains how this capability works.
Forge Studio allows you to inspect, execute, and verify the live implementation.
Capability Explorer
Browse the live capability catalog, supported execution surfaces, available Primitive Profiles, and execution metadata.
→ https://studio.forgepool.io/capability-explorer
Block Registry
Inspect the registered Primitive Profile, execution contract, block metadata, adapters, versions, and runtime characteristics.
→ https://studio.forgepool.io/studio/blocks-registry
Execute
Execute this capability using Forge Studio, the Execution API, or an MCP-compatible client.
Verify
Before interpreting the result, inspect:
- runtime evidence
- generated artifacts
- deterministic replay
- execution metadata
Trust should be established through independent verification rather than documentation alone.