Quant Swarm Audit — Execute Strategy Audit
audit.runRead-onlyIdempotent
Executes a reproducible Quant Swarm Audit job bundle with approval-token checks, idempotency deduplication, and a hard credit cap so retries and preflight refusals never trigger duplicate billing.
Instructions
Executes the Quant Swarm Audit and records a reproducible job bundle. Enforces approval_token verification, idempotency_key deduplication, and max_credits spend cap; never charges on preflight refusals.
Input Schema
| Name | Required | Description | Default |
|---|---|---|---|
| mode | No | Execution mode: 'local' (0 credits) or 'hosted' (20 credits). | local |
| max_credits | No | Hard spend cap in credits (0 for local runs; >=20 for hosted runs). | |
| trades_path | No | Optional trades path. | |
| trades_rows | No | Optional inline trades rows. | |
| preflight_id | No | Preflight ID returned by `audit.preflight`. | |
| random_seeds | No | Deterministic bootstrap seeds (must match the seeds used in `audit.quote`). | |
| returns_path | No | Optional returns path. | |
| returns_rows | No | Optional inline returns rows. | |
| strategy_name | No | Strategy identifier. | candidate_strategy |
| approval_token | Yes | Input-bound approval_token returned by `audit.quote`. | |
| benchmark_path | No | Optional benchmark path. | |
| benchmark_rows | No | Optional inline benchmark rows. | |
| idempotency_key | Yes | Caller-supplied idempotency key to prevent duplicate execution or billing on retries. | |
| cost_assumptions | No | Optional inline cost assumptions dict. | |
| features_manifest | No | Optional inline feature manifest dict. | |
| cost_assumptions_path | No | Optional cost assumptions JSON path. | |
| features_manifest_path | No | Optional feature manifest JSON path. |
Output Schema
| Name | Required | Description | Default |
|---|---|---|---|
| how | No | ||
| mode | No | ||
| tool | No | ||
| error | No | ||
| access | No | ||
| job_id | No | ||
| reason | No | ||
| status | No | ||
| metrics | No | ||
| eligible | No | ||
| findings | No | ||
| gbp_cost | No | ||
| job_id_a | No | ||
| job_id_b | No | ||
| next_step | No | ||
| strategy_a | No | ||
| strategy_b | No | ||
| html_report | No | ||
| max_credits | No | ||
| credits_cost | No | ||
| generated_at | No | ||
| input_sha256 | No | ||
| metrics_diff | No | ||
| new_warnings | No | ||
| preflight_id | No | ||
| random_seeds | No | ||
| has_benchmark | No | ||
| strategy_name | No | ||
| written_files | No | ||
| approval_token | No | ||
| execution_plan | No | ||
| input_sha256_a | No | ||
| input_sha256_b | No | ||
| trades_summary | No | ||
| artifact_hashes | No | ||
| coverage_matrix | No | ||
| credits_charged | No | ||
| idempotency_key | No | ||
| markdown_report | No | ||
| package_version | No | ||
| returns_summary | No | ||
| assessment_after | No | ||
| cost_assumptions | No | ||
| headline_metrics | No | ||
| not_tested_count | No | ||
| required_credits | No | ||
| assessment_before | No | ||
| idempotent_replay | No | ||
| request_access_at | No | ||
| resolved_findings | No | ||
| validation_errors | No | ||
| counts_by_category | No | ||
| overall_assessment | No | ||
| finding_transitions | No | ||
| remediation_verdict | No | ||
| has_features_manifest | No | ||
| not_investment_advice | No | ||
| outbound_data_preview | No | ||
| reproducibility_bundle | No | ||
| estimated_runtime_seconds | No |