cerebrixos-org/tuning-engines-cli
52Last commit Aug 15, 2026
tuning-engines-cli DSH plugin
Tuning Engines CLI & MCP Server provides a governed AI runtime for model, agent, skill, and MCP workflows. It includes a CLI for managing fine-tuning jobs, models, and orchestration starters, and a built-in MCP server with 60+ tools for AI assistants. The plugin integrates with DeepSeek Harness to capture telemetry and enforce governance.
How to install the tuning-engines-cli DSH plugin
dsh plugin --profile default add tuningengines-cliCopying does not run this command. Review the repository and version before installing the tuning-engines-cli DSH plugin.
tuning-engines-cli DSH plugin data source
tuning-engines-cli DSH plugin snapshot date: Aug 16, 2026
discovered
What the tuning-engines-cli DSH plugin can do
- Fine-tune open-source models using specialized agents (Cody for code autocomplete, SIERA for bug-fix).
- Manage training jobs: create, monitor status, list models, estimate costs.
- MCP server with 60+ tools for natural language interaction via AI assistants.
- Create governed orchestration starters for LangGraph, Temporal, and other frameworks.
- Route inference through an OpenAI-compatible API with RBAC, traffic policies, and traces.
Where the tuning-engines-cli DSH plugin fits
- Train a code autocomplete model on a GitHub repository using the Cody agent.
- Debug and fix errors by training a SIERA bug-fix specialist on your codebase.
- Monitor training progress and manage models from the command line.
- Connect an AI assistant (e.g., Claude) to fine-tune models and inspect traces via MCP.
Who the tuning-engines-cli DSH plugin is for
- Developers who need to fine-tune open-source models for code-specific tasks.
- Teams requiring governance, RBAC, and audit trails for AI inference and training workflows.
tuning-engines-cli DSH plugin limitations
- Requires a Tuning Engines account and API key (TE_API_KEY) to operate.
- MCP server intentionally does not expose internal proxy routes and refuses inference-key creation and raw secret-bearing mutations.
- Currently only code-focused agents (Cody, SIERA) are available; other domains (support, extraction, security) are planned but not yet released.
tuning-engines-cli DSH plugin: from the repository README
Quoted from the cerebrixos-org/tuning-engines-cli README, the upstream source of the tuning-engines-cli DSH plugin. Copyright remains with the original authors.
Govern every AI workflow through one API. **[Tuning Engines](https://tuningengines.com)** is a governed AI runtime for model, agent, skill, and MCP workflows. Route inference through one OpenAI-compatible API, apply RBAC and traffic policies, request approvals for high-risk actions, inspect traces and usage, and connect durable orchestration frameworks such as LangGraph and Temporal. The same CLI and MCP server also manage domain-specific fine-tuning of open-source models. ## Training Agents Tuning Engines uses specialized **agents** that control how your data is analyzed and converted into training data. Each agent produces a different kind of domain-specific fine-tuned model optimized for its use case. Current agents focus on code, with more coming for customer support, data extraction, security review, ops, and other domains. ### Cody (`code_repo`) — Code Autocomplete Agent Cody fine-tunes on your GitHub repo using QLoRA (4-bit quantized LoRA) via the Axolotl framework (HuggingFace Transformers + PEFT). It learns your codebase's patterns, naming conventions, and project structure to produce a fast, lightweight adapter optimized for real-time completions. **Best for:** code
Read the full READMERepository license: MIT
tuning-engines-cli DSH plugin questions
How do I install the Tuning Engines CLI DSH plugin?
Run the command `dsh plugin --profile default add tuningengines-cli` to install the plugin. You also need to set the `TE_API_KEY` environment variable with your Tuning Engines tenant token or inference key. After installation, the plugin will capture turn, model-step, and tool lifecycle telemetry.
What can I do with the MCP server?
The MCP server provides 60+ tools for AI assistants. You can fine-tune models, manage training jobs, run evaluations, check inference usage, inspect traces, review approvals, and manage non-secret tenant registry metadata. It works with any AI assistant that supports MCP, such as Claude Desktop, Claude Code, and VS Code extensions.
Do I need a Tuning Engines account to use the plugin?
Yes, you need a Tuning Engines account and an API key. You can sign up or log in using `te auth login` (opens a browser). The account is required to manage training jobs, store models, and enforce governance policies.
What base models are supported for fine-tuning?
The plugin supports a wide range of open-source models from 3B to 72B parameters, including Qwen, CodeLlama, DeepSeek, StarCoder, and Llama. You can check the full list in the README or use the `te jobs estimate` command to see supported models for a specific agent.
Can I use the plugin with Claude Code or Codex?
Yes, the plugin includes native telemetry hooks for Claude Code and Codex. You can install them with `te guard claude-code install` or `te guard codex install`. These hooks capture traces of your coding sessions and send pseudonymous references to Tuning Engines for governance.