> ## Documentation Index > Fetch the complete documentation index at: https://evalgate.com/docs/llms.txt > Use this file to discover all available pages before exploring further. # Create evaluations agent workflows > Requires scopes: eval:write ## OpenAPI ````yaml /api-reference/openapi.json post /api/evaluations/{id}/agent-workflows openapi: 3.1.0 info: title: EvalGate API version: 3.7.6 description: EvalGate API. See docs/api-contract.md for stability commitments. servers: - url: https://evalgate.com description: Production - url: http://localhost:3000 description: Local security: [] paths: /api/evaluations/{id}/agent-workflows: post: tags: - evaluations summary: Create evaluations agent workflows description: 'Requires scopes: eval:write' operationId: post_evaluations_id_agent_workflows parameters: - name: id in: path required: true schema: type: integer requestBody: required: true content: application/json: schema: $ref: '#/components/schemas/PostApiEvaluationsIdAgentWorkflowsRequest' responses: '201': description: Successful response content: application/json: schema: $ref: >- #/components/schemas/PostApiEvaluationsIdAgentWorkflowsResponse201ApplicationJson '400': $ref: '#/components/responses/ValidationError' '401': $ref: '#/components/responses/Unauthorized' '403': $ref: '#/components/responses/Forbidden' '404': $ref: '#/components/responses/NotFound' '409': $ref: '#/components/responses/Conflict' '429': $ref: '#/components/responses/RateLimited' '500': $ref: '#/components/responses/InternalError' security: - bearerAuth: - eval:write components: schemas: PostApiEvaluationsIdAgentWorkflowsRequest: type: object properties: runId: type: number caseCount: type: number includePassed: oneOf: - type: boolean enum: - false - type: boolean enum: - true persistCandidates: oneOf: - type: boolean enum: - false - type: boolean enum: - true persistArtifacts: oneOf: - type: boolean enum: - false - type: boolean enum: - true additionalProperties: false PostApiEvaluationsIdAgentWorkflowsResponse201ApplicationJson: type: object properties: coverage: type: object properties: totalCases: type: number agentWorkflowCases: type: number coverageRate: type: number workflowClassDistribution: type: object properties: approval_gated_mutation: type: number background_execution: type: number recovery_required: type: number policy_conflict: type: number handoff_heavy: type: number memory_sensitive: type: number parallel_tool_use: type: number additionalProperties: false required: - approval_gated_mutation - background_execution - recovery_required - policy_conflict - handoff_heavy - memory_sensitive - parallel_tool_use autonomyDistribution: type: object properties: manual: type: number default_approvals: type: number bypass_approvals: type: number autopilot: type: number additionalProperties: false required: - manual - default_approvals - bypass_approvals - autopilot generatedCandidateCount: type: number experimentRecipeCount: type: number gaps: type: array items: type: object properties: id: type: string workflowClass: oneOf: - type: string enum: - approval_gated_mutation - type: string enum: - background_execution - type: string enum: - recovery_required - type: string enum: - policy_conflict - type: string enum: - handoff_heavy - type: string enum: - memory_sensitive - type: string enum: - parallel_tool_use severity: oneOf: - type: string enum: - critical - type: string enum: - info - type: string enum: - warning description: type: string recommendation: type: string additionalProperties: false required: - id - workflowClass - severity - description - recommendation additionalProperties: false required: - totalCases - agentWorkflowCases - coverageRate - workflowClassDistribution - autonomyDistribution - generatedCandidateCount - experimentRecipeCount - gaps artifacts: type: array items: type: object properties: id: type: number kind: type: string title: type: string additionalProperties: false required: - id - kind - title cases: type: array items: type: object properties: evalCase: type: object properties: title: type: string id: type: string createdAt: type: string updatedAt: type: string tags: type: array items: type: string sourceTraceIds: type: array items: type: string expectedConstraints: type: array items: type: object properties: type: oneOf: - type: string enum: - custom - type: string enum: - contains - type: string enum: - json_schema - type: string enum: - not_contains - type: string enum: - matches_regex - type: string enum: - score_gte - type: string enum: - score_lte - type: string enum: - no_pii - type: string enum: - no_toxicity required: type: boolean value: $ref: '#/components/schemas/JsonValue' description: type: string additionalProperties: false required: - type - required quarantined: type: boolean evalCaseVersion: type: number frozenSnapshotRef: oneOf: - type: 'null' - type: string rubricRef: oneOf: - type: 'null' - type: string metricDAGRef: oneOf: - type: 'null' - type: string replayTier: oneOf: - type: 'null' - type: string enum: - A - type: string enum: - B - type: string enum: - C redactionProfileRef: oneOf: - type: 'null' - type: string mergedFromIds: type: array items: type: string agentWorkflow: type: object properties: version: type: number enum: - 1 autonomyLevel: oneOf: - type: string enum: - manual - type: string enum: - default_approvals - type: string enum: - bypass_approvals - type: string enum: - autopilot riskLevel: oneOf: - type: string enum: - low - type: string enum: - medium - type: string enum: - high - type: string enum: - critical workflowClass: oneOf: - type: string enum: - approval_gated_mutation - type: string enum: - background_execution - type: string enum: - recovery_required - type: string enum: - policy_conflict - type: string enum: - handoff_heavy - type: string enum: - memory_sensitive - type: string enum: - parallel_tool_use approvalMode: oneOf: - type: string enum: - human_review - type: string enum: - none - type: string enum: - policy_guided - type: string enum: - auto_with_audit memoryMode: oneOf: - type: string enum: - session - type: string enum: - none - type: string enum: - persistent - type: string enum: - mixed backgroundExecution: type: boolean requiredHooks: type: array items: oneOf: - type: string enum: - session_start - type: string enum: - pre_plan - type: string enum: - pre_tool - type: string enum: - post_tool - type: string enum: - pre_write - type: string enum: - pre_handoff - type: string enum: - on_error - type: string enum: - session_end requiresCheckpoint: type: boolean expectedHandoff: type: boolean activeTools: type: array items: type: string permissionScope: type: object properties: filesystem: type: array items: type: string network: oneOf: - type: string enum: - open - type: string enum: - restricted - type: string enum: - none externalSystems: type: array items: type: string mutatesState: oneOf: - type: boolean enum: - false - type: boolean enum: - true additionalProperties: false expectedFailureModes: type: array items: type: string scenario: type: string provenance: type: object properties: source: oneOf: - type: string enum: - manual - type: string enum: - evaluation_run - type: string enum: - coverage_gap - type: string enum: - trace_failure evaluationId: type: number runId: type: number testCaseId: type: number traceId: oneOf: - type: 'null' - type: string additionalProperties: false required: - source additionalProperties: false required: - version - autonomyLevel - riskLevel - workflowClass - approvalMode - memoryMode - backgroundExecution - requiredHooks - requiresCheckpoint - expectedHandoff - activeTools - permissionScope - expectedFailureModes - scenario additionalProperties: false required: - title - id - createdAt - updatedAt - tags - sourceTraceIds - expectedConstraints - quarantined - evalCaseVersion - frozenSnapshotRef - rubricRef - metricDAGRef - replayTier - redactionProfileRef - mergedFromIds quality: type: object properties: qualityScore: type: number usefulnessScore: type: number falsePositiveRisk: type: number compositeScore: type: number signals: type: array items: type: object properties: dimension: type: string score: type: number weight: type: number explanation: type: string additionalProperties: false required: - dimension - score - weight - explanation verdict: oneOf: - type: string enum: - low - type: string enum: - medium - type: string enum: - high - type: string enum: - reject recommendation: type: string additionalProperties: false required: - qualityScore - usefulnessScore - falsePositiveRisk - compositeScore - signals - verdict - recommendation rationale: type: string minimizedInput: type: object properties: userPrompt: type: string systemPrompt: oneOf: - type: 'null' - type: string activeTools: type: array items: type: string conversationContext: type: array items: type: object properties: role: oneOf: - type: string enum: - user - type: string enum: - system - type: string enum: - developer - type: string enum: - tool - type: string enum: - tool_result - type: string enum: - assistant content: type: string additionalProperties: false required: - role - content failureSpanId: oneOf: - type: 'null' - type: string failureOutput: oneOf: - type: 'null' - type: string metadata: type: object additionalProperties: $ref: '#/components/schemas/JsonValue' agentWorkflow: type: object properties: version: type: number enum: - 1 workflowClass: oneOf: - type: string enum: - approval_gated_mutation - type: string enum: - background_execution - type: string enum: - recovery_required - type: string enum: - policy_conflict - type: string enum: - handoff_heavy - type: string enum: - memory_sensitive - type: string enum: - parallel_tool_use autonomyLevel: oneOf: - type: string enum: - manual - type: string enum: - default_approvals - type: string enum: - bypass_approvals - type: string enum: - autopilot approvalMode: oneOf: - type: string enum: - human_review - type: string enum: - none - type: string enum: - policy_guided - type: string enum: - auto_with_audit riskLevel: oneOf: - type: string enum: - low - type: string enum: - medium - type: string enum: - high - type: string enum: - critical memoryMode: oneOf: - type: string enum: - session - type: string enum: - none - type: string enum: - persistent - type: string enum: - mixed backgroundExecution: type: boolean requiresCheckpoint: type: boolean expectedHandoff: type: boolean requiredHooks: type: array items: oneOf: - type: string enum: - session_start - type: string enum: - pre_plan - type: string enum: - pre_tool - type: string enum: - post_tool - type: string enum: - pre_write - type: string enum: - pre_handoff - type: string enum: - on_error - type: string enum: - session_end activeTools: type: array items: type: string permissionScope: type: object properties: filesystem: type: array items: type: string network: oneOf: - type: string enum: - open - type: string enum: - restricted - type: string enum: - none externalSystems: type: array items: type: string mutatesState: oneOf: - type: boolean enum: - false - type: boolean enum: - true additionalProperties: false expectedFailureModes: type: array items: type: string scenario: type: string provenance: type: object properties: source: oneOf: - type: string enum: - manual - type: string enum: - evaluation_run - type: string enum: - coverage_gap - type: string enum: - trace_failure evaluationId: type: number runId: type: number testCaseId: type: number traceId: oneOf: - type: 'null' - type: string additionalProperties: false required: - source additionalProperties: false required: - version - workflowClass - autonomyLevel - approvalMode - riskLevel - memoryMode - backgroundExecution - requiresCheckpoint - expectedHandoff - requiredHooks - activeTools - permissionScope - expectedFailureModes - scenario workflowExperimentRecipe: oneOf: - type: 'null' - type: object properties: version: type: number enum: - 1 title: type: string objective: type: string workflowClass: oneOf: - type: string enum: - approval_gated_mutation - type: string enum: - background_execution - type: string enum: - recovery_required - type: string enum: - policy_conflict - type: string enum: - handoff_heavy - type: string enum: - memory_sensitive - type: string enum: - parallel_tool_use source: oneOf: - type: string enum: - coverage_gap - type: string enum: - failure linkedFailureModes: type: array items: type: string linkedTestCaseIds: type: array items: type: number comparedVariants: type: array items: type: object properties: id: type: string label: type: string changes: type: array items: type: string autonomyLevel: oneOf: - type: string enum: - manual - type: string enum: - default_approvals - type: string enum: - bypass_approvals - type: string enum: - autopilot approvalMode: oneOf: - type: string enum: - human_review - type: string enum: - none - type: string enum: - policy_guided - type: string enum: - auto_with_audit memoryMode: oneOf: - type: string enum: - session - type: string enum: - none - type: string enum: - persistent - type: string enum: - mixed hooks: type: array items: oneOf: - type: string enum: - session_start - type: string enum: - pre_plan - type: string enum: - pre_tool - type: string enum: - post_tool - type: string enum: - pre_write - type: string enum: - pre_handoff - type: string enum: - on_error - type: string enum: - session_end expectedImpact: type: string additionalProperties: false required: - id - label - changes - expectedImpact successMetrics: type: array items: type: string guardrails: type: array items: type: string recommendedIterations: type: number rationale: type: string additionalProperties: false required: - version - title - objective - workflowClass - source - linkedFailureModes - linkedTestCaseIds - comparedVariants - successMetrics - guardrails - recommendedIterations - rationale additionalProperties: false required: - userPrompt - systemPrompt - activeTools - conversationContext - failureSpanId - failureOutput - metadata recipe: type: object properties: version: type: number enum: - 1 title: type: string objective: type: string workflowClass: oneOf: - type: string enum: - approval_gated_mutation - type: string enum: - background_execution - type: string enum: - recovery_required - type: string enum: - policy_conflict - type: string enum: - handoff_heavy - type: string enum: - memory_sensitive - type: string enum: - parallel_tool_use source: oneOf: - type: string enum: - coverage_gap - type: string enum: - failure linkedFailureModes: type: array items: type: string linkedTestCaseIds: type: array items: type: number comparedVariants: type: array items: type: object properties: id: type: string label: type: string changes: type: array items: type: string autonomyLevel: oneOf: - type: string enum: - manual - type: string enum: - default_approvals - type: string enum: - bypass_approvals - type: string enum: - autopilot approvalMode: oneOf: - type: string enum: - human_review - type: string enum: - none - type: string enum: - policy_guided - type: string enum: - auto_with_audit memoryMode: oneOf: - type: string enum: - session - type: string enum: - none - type: string enum: - persistent - type: string enum: - mixed hooks: type: array items: oneOf: - type: string enum: - session_start - type: string enum: - pre_plan - type: string enum: - pre_tool - type: string enum: - post_tool - type: string enum: - pre_write - type: string enum: - pre_handoff - type: string enum: - on_error - type: string enum: - session_end expectedImpact: type: string additionalProperties: false required: - id - label - changes - expectedImpact successMetrics: type: array items: type: string guardrails: type: array items: type: string recommendedIterations: type: number rationale: type: string additionalProperties: false required: - version - title - objective - workflowClass - source - linkedFailureModes - linkedTestCaseIds - comparedVariants - successMetrics - guardrails - recommendedIterations - rationale source: oneOf: - type: string enum: - coverage_gap - type: string enum: - run_failure testCaseId: type: number additionalProperties: false required: - evalCase - quality - rationale - minimizedInput - recipe - source createdCandidateIds: type: array items: type: number runId: oneOf: - type: 'null' - type: number additionalProperties: false required: - coverage - artifacts - cases - createdCandidateIds - runId JsonValue: oneOf: - type: 'null' - type: boolean - type: number - type: string - type: array items: $ref: '#/components/schemas/JsonValue' - type: object additionalProperties: $ref: '#/components/schemas/JsonValue' ApiError: type: object properties: error: type: object properties: code: type: string enum: - UNAUTHORIZED - FORBIDDEN - NOT_FOUND - VALIDATION_ERROR - RATE_LIMITED - CONFLICT - INTERNAL_ERROR - SERVICE_UNAVAILABLE - QUOTA_EXCEEDED - NO_ORG_MEMBERSHIP - CROSS_ORG_REFERENCE - INCOMPLETE - MALFORMED message: type: string details: {} requestId: type: string format: uuid required: - code - message required: - error responses: ValidationError: description: Validation error content: application/json: schema: $ref: '#/components/schemas/ApiError' Unauthorized: description: Unauthorized content: application/json: schema: $ref: '#/components/schemas/ApiError' Forbidden: description: Forbidden content: application/json: schema: $ref: '#/components/schemas/ApiError' NotFound: description: Resource not found content: application/json: schema: $ref: '#/components/schemas/ApiError' Conflict: description: Conflict content: application/json: schema: $ref: '#/components/schemas/ApiError' RateLimited: description: Rate limit exceeded content: application/json: schema: $ref: '#/components/schemas/ApiError' InternalError: description: Internal server error content: application/json: schema: $ref: '#/components/schemas/ApiError' securitySchemes: bearerAuth: type: http scheme: bearer bearerFormat: API Key or Session Token ````