Knuckles-Team Portfolio Stress Test Workflow
CONCEPT:EE-011
Aggregates return curves, runs historical crash scenarios, computes rate shock exposures, and compiles risk audits.
Steps
Step 1: Returns Data Crawlers
Agent: data-fetcher
Tools: graph_query, sx_search
Gathers historical holdings and benchmarks returns data.
Expected: historical-stress-pricing-data
Step 2: Scenario Crash Simulator [depends_on: returns-data-crawlers]
Agent: compute-engine
Tools: graph_analyze
Simulates returns during historical financial crisis events.
Expected: crisis-returns-scenarios
Step 3: Interest Rate Shock [depends_on: returns-data-crawlers]
Agent: risk-assessor
Tools: graph_query, graph_analyze
Applies simulated treasury yield spikes to active pricing metrics.
Expected: interest-rate-shock-scenarios
Step 4: Stress Aggregation [depends_on: scenario-crash-simulator, interest-rate-shock]
Agent: report-generator
Tools: graph_write, document_tools
Aggregates stress-test drawdowns into a compliance card.
Expected: stress-metrics-scorecard
Step 5: KG Persistence [depends_on: stress-aggregation]
Agent: report-generator
Tools: graph_write
Persist workflow results as nodes and edges in the Knowledge Graph. Create appropriate typed nodes with metadata and link to existing domain entities.
Output
- Portfolio Stress Test results persisted in KG
- Structured report (MD/PDF)
- Audit trail with timestamps and agent attributions
Execution
Run this workflow as a dependency-ordered DAG. Steps with no unmet depends_on run in parallel; dependents run after their prerequisites complete.
- Run first (in parallel): Step 1 — Returns Data Crawlers
- After level 0: Step 2 — Scenario Crash Simulator; Step 3 — Interest Rate Shock
- After level 1: Step 4 — Stress Aggregation
- After level 2: Step 5 — KG Persistence
Execution: If graph-os is reachable, offload the whole DAG via graph_orchestrate action=execute_workflow (or the kg-delegate skill) for true parallel/swarm execution. Otherwise execute the steps natively in dependency order: run steps with no unmet depends_on in parallel, then their dependents.