Mistral Port

Purpose

This is the reference integration for Harbor http_api Action Model authoring for Mistral.

Use it when:

• building Mistral actions in Harbor
• extending the Dock / Hub Mistral entry
• teaching Codex or OpenClaw how to map a model API into Harbor Ports and Harbor Actions

Harbor Port worksheet

Product

• Product: Mistral API
• Publisher slug: mistral
• Publisher name: Mistral AI
• Category: AI
• Tags: mistral, ai, llm, chat, embeddings

Safe Harbor Port metadata

• Kind: http_api
• Label: Mistral API
• Description: Mistral API connection managed locally through Harbor.
• Base URL: https://api.mistral.ai
• Auth mode: header_token
• Auth header name: Authorization
• Auth token prefix: Bearer

Hidden local config

• authToken

Harbor stores the token locally only. Dock / Hub manifests must never contain it.

Harbor action design notes

Mistral is a good Harbor fit when we stay on fixed v1 routes, keep model execution inside explicit JSON request bodies, and treat prompt-bearing calls as approval-friendly by default.

Good Harbor Mistral actions:

• stable reads like GET /v1/models
• fixed prompt-bearing routes like POST /v1/chat/completions
• fixed embedding routes like POST /v1/embeddings
• requestBodyMode: "none" for reads
• requestBodyMode: "json" for chat and embeddings
• resultMode: "json_summary" for all starter actions

Avoid for now:

• streaming-specific runtime handling
• OCR, transcription, and file-upload specific flows
• broader platform and workspace administration surfaces
• approval-free prompt execution

Starter Mistral actions

Automatic read actions

• list-models -> GET /v1/models

Approval-friendly write actions

• chat-complete -> POST /v1/chat/completions
• embed-text -> POST /v1/embeddings

Why these actions

These routes fit Harbor's Action Model well because they:

• stay under one fixed Mistral API origin
• use stable documented v1 paths
• keep prompts and embedding payloads in explicit JSON bodies
• map cleanly to named, reviewable AI capabilities
• keep hidden auth entirely local to Harbor Node

Suggested operator workflow

1. Import the Mistral Dock entry from Dock / Hub.
2. Configure the local Mistral API key in Harbor Node.
3. Validate list-models first to confirm local auth and account setup.
4. Keep chat-complete and embed-text on require_approval unless there is a deliberate local policy reason not to.
5. Expand only after validating that the prompt and usage patterns fit your Harbor Guard policy.

Suggested agent workflow

1. Inspect the existing Mistral Harbor Port.
2. Inspect live actions and drafts.
3. Propose only fixed-route additions that match the current auth and JSON body model.
4. Validate and test drafts.
5. Request publish instead of self-publishing unless Harbor explicitly allows it.

Agents should never:

• ask Harbor for the stored API key
• widen Mistral into a generic arbitrary HTTP proxy
• bypass approval for prompt-bearing actions without operator intent

Current update strategy

For now, Mistral updates are manual:

• update the Dock / Hub integration document in apps/hub/catalog/integrations/mistral/mistral-api.json
• bump the manifest version when the action set changes materially
• re-import into Harbor or create new drafts from the updated Dock entry

Official reference starting points

• Mistral API docs: https://docs.mistral.ai/api
• Mistral quickstart: https://docs.mistral.ai/getting-started/quickstart/
• Mistral chat completions: https://docs.mistral.ai/api/
• Mistral model docs: https://docs.mistral.ai/models