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

# Read the account's script-execution opt-in

> On unless an admin turned it off: an account with no stored document reports `enabled: true`. Readable by **any member** — knowing whether Abbie may run a skill's code is not privileged. A read failure degrades to `enabled: false` rather than an error: reporting a security switch as open on the strength of a failed read would be the wrong direction.

This is the middle of three consent layers in front of script execution: the deployment-wide `SKILL_EXEC_ENABLED` flag, this per-account opt-in, and the per-skill `exec_approved` bit on `PUT /hub/coworker/skill`. All three must be open. Subject: `microstrate.hub.get.coworker-skill-exec`.



## OpenAPI

````yaml /api-reference/endpoint/abbie/openapi.json get /hub/coworker/skill-exec
openapi: 3.1.0
info:
  title: Abbie API (beta)
  version: 1.0.0
  description: >
    **Beta.** This API is in beta: operations and fields may change without a
    deprecation period. Do not depend on undocumented fields. It covers the
    subset of Abbie's API used by the Quiva coworker MCP server: skills, todos,
    memory, profile, corrections, environment and models.


    HTTP surface of hub-service for Abbie/Coworker configuration, agent rapport
    & context, chat-session sharing, and MCP registration.


    ## Transport and response envelope


    Every operation here is a gateway mapping onto a Bellerophon/kraken subject
    of the form `microstrate.hub.<method>.<endpoint-slug>` (see
    `hub-service/service/service.go` and `docs/abbie-gateway-mappings.md`).
    Handlers reply on the mesh with an envelope `{"status_code": <int>, "body":
    <object>}`; the gateway unwraps it, so `status_code` becomes the HTTP status
    and `body` becomes the HTTP response body. **The schemas below describe the
    unwrapped `body`.**


    Errors are always `{"error": "<message>"}` with the matching HTTP status.


    ## Parameter conventions


    The gateway forwards URL parameters to the handler as request headers:


    - `x-param-query-<name>` — a query-string parameter (`?name=value`)

    - `x-param-<name>` — a path parameter


    Every parameter documented as `in: query` or `in: path` below corresponds to
    one of those headers, and the header suffix is named in each parameter's
    description where it is not obvious.


    ## Authentication


    All operations require the caller's bearer token; the account id, user id
    and user public key are always taken from the JWT and never from the URL or
    body. Endpoints that additionally require an account **admin** role are
    marked in their description — a failed role lookup answers **502**, never
    403 and never a pass.


    ## Two path notes


    - `PATCH /hub/agent-session-visibility/{session_id}` — the handler reads the
    path parameter as `x-param-session_id`, matching its sibling `PATCH
    /hub/agent-session/{session_id}`.

    - `GET|POST|PUT /hub/agent/user-context/{id}` — the handler requires the
    **agent id** as `x-param-id`; the frontend calls
    `/hub/agent/user-context/<agentId>`.
servers:
  - url: https://api.quiva.ai
    description: Production API server
security:
  - bearerAuth: []
  - apiKeyAuth: []
tags:
  - name: Skills
  - name: Todos
  - name: Memory
  - name: Profile and personality
  - name: Corrections
  - name: Environment
  - name: Models
paths:
  /hub/coworker/skill-exec:
    get:
      tags:
        - Skills
      summary: Read the account's script-execution opt-in
      description: >-
        On unless an admin turned it off: an account with no stored document
        reports `enabled: true`. Readable by **any member** — knowing whether
        Abbie may run a skill's code is not privileged. A read failure degrades
        to `enabled: false` rather than an error: reporting a security switch as
        open on the strength of a failed read would be the wrong direction.


        This is the middle of three consent layers in front of script execution:
        the deployment-wide `SKILL_EXEC_ENABLED` flag, this per-account opt-in,
        and the per-skill `exec_approved` bit on `PUT /hub/coworker/skill`. All
        three must be open. Subject: `microstrate.hub.get.coworker-skill-exec`.
      operationId: getCoworkerSkillExec
      responses:
        '200':
          description: Consent state
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/SkillExecConsent'
        '401':
          $ref: '#/components/responses/Unauthorized'
components:
  schemas:
    SkillExecConsent:
      type: object
      properties:
        enabled:
          type: boolean
        enabled_by:
          type: string
          description: Accounts user id — the audit trail
        enabled_at:
          type: string
          format: date-time
    Error:
      type: object
      description: The unwrapped error body. The HTTP status carries the code.
      properties:
        error:
          type: string
      required:
        - error
  responses:
    Unauthorized:
      description: No account (or user) identity in the auth context
      content:
        application/json:
          schema:
            $ref: '#/components/schemas/Error'
  securitySchemes:
    bearerAuth:
      type: http
      scheme: bearer
      bearerFormat: JWT
    apiKeyAuth:
      type: apiKey
      in: header
      name: x-api-key

````

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