> ## Documentation Index
> Fetch the complete documentation index at: https://docs.infercrane.com/llms.txt
> Use this file to discover all available pages before exploring further.

# List revision-bound semantic quality evidence and per-eval PASS, WATCH, or BLOCK verdicts without prompt or output content



## OpenAPI

````yaml /openapi.json get /api/v1/deployments/{name}/quality-evidence
openapi: 3.1.0
info:
  description: >-
    Durable, authenticated control-plane operations. Mutation responses may
    outlive the requesting client.
  title: InferCrane Control API
  version: 1.0.0
servers:
  - description: Local InferCrane API and gateway
    url: http://127.0.0.1:18000
security: []
tags:
  - name: Admission
  - name: Adoption
  - name: Agents
  - name: Alerts
  - name: Async inference
  - name: Audit
  - name: Autopilot
  - name: Benchmarks
  - name: Billing
  - name: Billing administration
  - name: Burst Guard
  - name: Capacity intelligence
  - name: Compute
  - name: Context Passport
  - name: Deployments
  - name: Diagnostics
  - name: Endpoints
  - name: External capacity
  - name: FinOps
  - name: Identity
  - name: Inference
  - name: Inference Lab
  - name: Inference decisions
  - name: Inference optimization
  - name: Infrastructure
  - name: Marketplace administration
  - name: Model API administration
  - name: Model APIs
  - name: Model artifacts
  - name: Model catalog
  - name: Monitoring
  - name: Notifications
  - name: Operations
  - name: Planning
  - name: Providers
  - name: Recipes
  - name: Release Guard
  - name: Release evidence
  - name: Replay
  - name: Revisions
  - name: Routing
  - name: Sandboxes
  - name: Scaling
  - name: Secrets
  - name: System
  - name: Targets
  - name: Training lineage
paths:
  /api/v1/deployments/{name}/quality-evidence:
    get:
      tags:
        - Release evidence
      summary: >-
        List revision-bound semantic quality evidence and per-eval PASS, WATCH,
        or BLOCK verdicts without prompt or output content
      operationId: listQualityEvidence
      parameters:
        - in: path
          name: name
          required: true
          schema:
            minLength: 1
            type: string
      responses:
        '200':
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/QualityEvidenceList'
          description: Success
        default:
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorEnvelope'
          description: Typed API error
      security:
        - bearerAuth: []
components:
  schemas:
    QualityEvidenceList:
      additionalProperties: false
      properties:
        content_recorded:
          const: false
          type: boolean
        data:
          items:
            additionalProperties: true
            type: object
          type: array
        verdicts:
          $ref: '#/components/schemas/QualityVerdicts'
      required:
        - data
        - verdicts
        - content_recorded
      type: object
    ErrorEnvelope:
      properties:
        error:
          $ref: '#/components/schemas/Error'
      required:
        - error
      type: object
    QualityVerdicts:
      additionalProperties: false
      description: >-
        Verdicts for the current candidate against the active revision under
        persisted Release Guard policy. Only state compared carries verdict,
        comparison_mode, reference, candidate, and evals. The stored Release
        Guard evaluation remains the decision of record.
      properties:
        candidate:
          $ref: '#/components/schemas/QualitySubject'
        comparison_mode:
          enum:
            - threshold
            - bootstrap
          type: string
        decision_authority:
          type: string
        evals:
          items:
            $ref: '#/components/schemas/QualityEvalVerdict'
          type: array
        max_regression_percent:
          type: number
        minimum_score:
          type: number
        reference:
          $ref: '#/components/schemas/QualitySubject'
        state:
          enum:
            - compared
            - no_candidate
            - not_gated
            - unavailable
          type: string
        verdict:
          description: >-
            Per-model verdict: the most severe eval verdict. UNKNOWN when
            evidence is missing or not comparable.
          enum:
            - PASS
            - WATCH
            - BLOCK
            - UNKNOWN
          type: string
        watch_regression_percent:
          type: number
      required:
        - state
        - decision_authority
      type: object
    Error:
      properties:
        category:
          type: string
        code:
          type: string
        message:
          type: string
        remediation:
          type: string
        request_id:
          type: string
        retryable:
          type: boolean
      required:
        - code
        - message
      type: object
    QualitySubject:
      additionalProperties: false
      properties:
        model:
          description: >-
            Resolved immutable model identity, else the model the revision
            requested; empty when neither is recorded.
          type: string
        revision_id:
          type: string
      required:
        - revision_id
        - model
      type: object
    QualityEvalVerdict:
      additionalProperties: false
      properties:
        bootstrap:
          additionalProperties: true
          type: object
        candidate_evidence_id:
          type: string
        candidate_passed:
          type: boolean
        candidate_score:
          type: number
        code:
          type: string
        delta:
          description: Candidate score minus reference score; negative is a drop.
          type: number
        evaluator:
          type: string
        evaluator_version:
          type: string
        follow_up:
          description: Operator action recorded for a WATCH verdict.
          type: string
        limit:
          description: Persisted threshold that decided a WATCH or BLOCK verdict.
          type: number
        message:
          type: string
        reference_evidence_id:
          type: string
        reference_score:
          type: number
        regression_percent:
          description: Drop relative to the reference score.
          type: number
        suite:
          type: string
        suite_version:
          type: string
        verdict:
          enum:
            - PASS
            - WATCH
            - BLOCK
            - UNKNOWN
          type: string
      required:
        - suite
        - suite_version
        - evaluator
        - evaluator_version
        - verdict
        - code
        - message
        - candidate_evidence_id
        - candidate_score
        - candidate_passed
      type: object
  securitySchemes:
    bearerAuth:
      scheme: bearer
      type: http

````

This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.