mta-lightspeed

name	mta-lightspeed
metadata	{"author":"rhoai3-coding-demo","version":"1.0.0","platform-family":"mta","platform-baseline":"repo","ocp-baseline":"repo","skill-group":"Application Modernization"}
description	Use when configuring or using Red Hat Developer Lightspeed for MTA to modernize applications with AI assistance, including LLM provider configuration, code migration suggestions, and AI-assisted analysis. Do NOT use for installing MTA (use mta-install) or non-AI tool usage (use mta-cli/mta-ui).

MTA Developer Lightspeed

Use this skill for configuring and using Red Hat Developer Lightspeed for Migration Toolkit for Applications (MTA) 8.2 — the AI-assisted code migration component that integrates LLMs with MTA static analysis.

Source Grounding

Read references/source-capture.md before using product configuration details. Official Red Hat docs are product authority. This skill covers LLM provider configuration, Solution Server setup, Agent AI mode, VS Code extension settings, and AI-assisted code resolution workflows.

Support Posture

Important: Developer Lightspeed for MTA and the Solution Server are Technology Preview features. They are not supported with Red Hat production SLAs and might not be functionally complete. Do not use them in production.

Concepts

Developer Lightspeed for MTA

Starting from MTA 8.0.0, the VS Code extension integrates with LLMs through the Developer Lightspeed component. It applies LLM-driven code changes to resolve issues found through static code analysis of Java applications, using Retrieval Augmented Generation (RAG) for context-based resolutions.

Solution Server

An optional component that builds collective memory of source code changes across analyses. It stores solved examples, computes success metrics (confidence levels), and improves migration hints over successive migration waves. Requires the LLM proxy service and a 5 Gi RWO persistent volume.

Agent AI

An automated analysis mode that plans context, selects sub-agents, and iterates through code fixes. Accepts changes, recompiles, and re-analyzes in a loop until all issues are resolved or a maximum of two attempts per issue is reached.

Supported LLM Providers

Provider (Tackle CR value)	Example models
OpenShift AI platform	OpenAI-compatible models deployed on-cluster
OpenAI (openai)	gpt-4, gpt-4o, gpt-4o-mini, gpt-3.5-turbo
Azure OpenAI (azure_openai)	gpt-4, gpt-35-turbo
Amazon Bedrock (bedrock)	anthropic.claude-3-5-sonnet-*, meta.llama3-1-70b-instruct-*
Google Gemini (google)	gemini-2.0-flash-exp, gemini-1.5-pro
Ollama (ollama)	llama3.1, codellama, mistral

Workflow

1. Read references/source-capture.md and confirm the product baseline.
2. Read references/official-doc-extraction.md for detailed procedures.
3. Select a deployment workflow:
   • Agent AI without LLM proxy — direct LLM connection via provider-settings.yaml.
   • LLM proxy with optional Solution Server — centralised key management via Tackle CR proxy service.
4. Configure LLM API key Secret in openshift-mta namespace.
5. Configure Tackle CR with provider and model settings.
6. Configure the VS Code extension (provider-settings.yaml, GenAI settings).
7. Run analysis and request AI-assisted code resolutions.

LLM API Key Secret

oc create secret generic kai-api-keys -n openshift-mta \
  --from-literal=OPENAI_API_KEY='<YOUR_KEY>'

Provider-specific variants exist for Azure, Bedrock, Google, and OpenShift AI. See references/official-doc-extraction.md.

Tackle CR Configuration

kind: Tackle
apiVersion: tackle.konveyor.io/v1alpha1
metadata:
  name: mta
  namespace: openshift-mta
spec:
  kai_llm_proxy_enabled: true
  kai_solution_server_enabled: true
  kai_llm_provider: <provider-name>
  kai_llm_model: <model-name>

Validation

oc get deploy,svc -n openshift-mta | grep -E 'kai-(api|db|importer)'

Data Privacy

Code snippets are transmitted intact to the configured LLM. No automatic sanitisation or redaction is performed. Organisations with strict IP protection should deploy self-managed models on OpenShift AI or RHEL AI and point Developer Lightspeed at the private endpoint.

VS Code Extension Settings

Key settings in Extensions > MTA:

Setting	Purpose
Gen AI: Enabled	Enable AI-assisted code fixes (default: true)
Gen AI: Agent mode	Enable automated agentic analysis loop
Gen AI: Excluded diagnostic sources	Skip specific diagnostic sources in Agent AI
Cache directory	Store cached LLM responses
Trace enabled / Trace directory	Trace MTA–LLM communication for debugging

Debugging

• Extension logs: Developer: Open Extension Logs Folder > redhat.mta-vscode-extension/extension.log.
• Output panel: select Red Hat Developer Lightspeed for MTA.
• Archive logs: MTA: Generate Debug Archive in Command Palette.

Related Skills

• mta-cli — CLI-based application analysis and migration.
• mta-ui — web UI analysis and assessment workflows.
• mta-install — MTA Operator installation and Tackle CR management.
• mta-vscode — VS Code extension non-AI features.
• mta-intellij — IntelliJ IDEA plugin workflows.

References

• references/source-capture.md
• references/official-doc-extraction.md