Lakera · Capability

Lakera Guard — Content Screening

Lakera Guard — Content Screening. 2 operations. Lead operation: Screen Content For Threats. Self-contained Naftiko capability covering the Lakera Guard runtime screening surface.

Lakera Guard — Content Screening is a Naftiko capability published by Lakera, one of 2 capabilities the APIs.io network indexes for this provider. It bundles 2 operations across the POST method rooted at /v2/guard.

The capability includes 2 read-only operations. Lead operation: Screen LLM messages for prompt attacks, data leakage, content violations, PII, and malicious links. Can be deployed as a REST endpoint, MCP tool, or Agent Skill via Naftiko.

Tagged areas include Lakera, Guard, AI Security, and Prompt Injection.

Run with Naftiko LakeraGuardAI SecurityPrompt Injection

What You Can Do

POST
Screencontent — Screen Content For Threats
/v2/guard
POST
Getresults — Get Detector Confidence Results
/v2/guard/results

MCP Tools

lakera-screen-content

Screen LLM messages for prompt attacks, data leakage, content violations, PII, and malicious links.

read-only idempotent
lakera-get-detector-results

Retrieve Lakera detector confidence levels for offline threshold tuning and quality monitoring.

read-only idempotent

Capability Spec

guard-screening.yaml Raw ↑ 
naftiko: 1.0.0-alpha2
info:
  label: Lakera Guard — Content Screening
  description: 'Lakera Guard — Content Screening. 2 operations. Lead operation: Screen Content For Threats. Self-contained
    Naftiko capability covering the Lakera Guard runtime screening surface.'
  tags:
  - Lakera
  - Guard
  - AI Security
  - Prompt Injection
  created: '2026-05-25'
  modified: '2026-05-25'
binds:
- namespace: env
  keys:
    LAKERA_GUARD_API_KEY: LAKERA_GUARD_API_KEY
capability:
  consumes:
  - type: http
    namespace: guard-screening
    baseUri: https://api.lakera.ai/v2
    description: Lakera Guard content screening capability. Self-contained, no shared references.
    resources:
    - name: guard
      path: /guard
      operations:
      - name: screenContent
        method: POST
        description: Screen Content For Threats
        outputRawFormat: json
        outputParameters:
        - name: result
          type: object
          value: $.
        inputParameters:
        - name: body
          in: body
          type: object
          description: Request body (JSON) with messages, optional project_id, policy_id, breakdown, payload.
          required: true
    - name: guard-results
      path: /guard/results
      operations:
      - name: getResults
        method: POST
        description: Get Detector Confidence Results
        outputRawFormat: json
        outputParameters:
        - name: result
          type: object
          value: $.
        inputParameters:
        - name: body
          in: body
          type: object
          description: Request body (JSON) with messages and optional project_id.
          required: true
    authentication:
      type: apikey
      key: Authorization
      value: 'Bearer {{env.LAKERA_GUARD_API_KEY}}'
      placement: header
  exposes:
  - type: rest
    namespace: guard-screening-rest
    port: 8080
    description: REST adapter for Lakera Guard content screening. One Spectral-compliant resource per consumed
      operation, prefixed with /v2.
    resources:
    - path: /v2/guard
      name: guard
      description: REST surface for /guard.
      operations:
      - method: POST
        name: screenContent
        description: Screen Content For Threats
        call: guard-screening.screenContent
        with:
          body: rest.body
        outputParameters:
        - type: object
          mapping: $.
    - path: /v2/guard/results
      name: guard-results
      description: REST surface for /guard/results.
      operations:
      - method: POST
        name: getResults
        description: Get Detector Confidence Results
        call: guard-screening.getResults
        with:
          body: rest.body
        outputParameters:
        - type: object
          mapping: $.
  - type: mcp
    namespace: guard-screening-mcp
    port: 9090
    transport: http
    description: MCP adapter for Lakera Guard content screening. One tool per consumed operation.
    tools:
    - name: lakera-screen-content
      description: Screen LLM messages for prompt attacks, data leakage, content violations, PII, and malicious links.
      hints:
        readOnly: true
        destructive: false
        idempotent: true
      call: guard-screening.screenContent
      with:
        body: tools.body
      outputParameters:
      - type: object
        mapping: $.
    - name: lakera-get-detector-results
      description: Retrieve Lakera detector confidence levels for offline threshold tuning and quality monitoring.
      hints:
        readOnly: true
        destructive: false
        idempotent: true
      call: guard-screening.getResults
      with:
        body: tools.body
      outputParameters:
      - type: object
        mapping: $.