CodeInsights

Coding Insights

Coding Insights agents provide in-depth analysis and answers across single or multiple code repositories. These agents help you explore, validate, and enhance codebases by answering questions, reviewing pull requests, and offering improvement suggestions — all powered by repository-aware intelligence.

1. Single Repo Code Lens

Enables deep analysis and Q&A over a single MuleSoft code repository using natural language — ideal for reviewing code quality, diagnosing issues, and exploring implementation logic.

• Natural Language Interface: Ask contextual questions about a specific repo and get direct answers from the CurieTech AI agent.
• Smart Code Understanding: The agent analyzes flows, APIs, commits, and connector configurations to generate responses.
• Live Repo Context: Connects to your Git repo or uploaded project to reference actual implementation logic and metadata.

Supported Input Modes:

• With Repository: Select a GitHub repository and branch to query directly.
• Upload from Computer: Upload a local MuleSoft project folder to interact without Git access.

Additional Features:

• "Ask a Question" Interface: Use the built-in prompt box to ask any code-related question about the selected repo.
• Deep Context Awareness: Agent considers flows, configurations, error handlers, and connector versions when answering.

Example Questions You Can Ask:

• "What APIs are present in this repository?"
• "Which flows have missing error handling?"
• "Write a Karate test scenario for the /xyz endpoint."
• "What test cases should I write to validate PR #123?"
• "Does this code follow MuleSoft best practices?"
• "What Salesforce connector version is used and is it the latest?"
• "This DWL script isn’t working — explain what’s wrong: <paste script>"

This agent is ideal for engineers looking to review, debug, or understand their MuleSoft codebase — without manually reading through every line.

2. Multi Repo Code Lens

Extends the capabilities of Single Repo Code Lens by allowing you to query and analyze up to 8 related MuleSoft repositories simultaneously — perfect for understanding cross-repo dependencies, ensuring design consistency, and performing architectural audits.

• Cross-Repo Intelligence: Evaluate patterns, inconsistencies, or dependencies across multiple services or layers.
• Ideal for Platform Architects: Gain insights into how System, Process, and Experience APIs interconnect or diverge.
• Connector and Flow Mapping: Understand shared logic, component usage, and distributed architecture in large MuleSoft ecosystems.

Supported Input Mode:

• With Repository (Multi-Select): Connect a GitHub branch and select up to 8 repositories to query at once.

Additional Features:

• "Ask a Question" Interface: Use the prompt box to ask cross-repository questions.
• Smart Cross-Referencing: Agent traces logic, connector usage, and configuration relationships across services.

Example Questions You Can Ask:

• "I created pull request #123 in the System API repo — do I need to change anything in the Process API?"
• "For pull request #123 in the System API, write a Karate test scenario in the Karate test repo."
• "Is the createOrder flow in Repo A calling any APIs from Repo B?"
• "Do all these APIs implement consistent error handling for HTTP 500 errors?"
• "Which repos are still using deprecated Mule components?"
• "Show me an example of how the S3 connector is used across any of these repositories."

This agent is best used for orchestrated platform reviews, dependency analysis, and aligning standards across multiple MuleSoft services.

3. Code Review Lens

Performs in-depth reviews of pull requests for a selected MuleSoft repository — identifying issues, validating best practices, and offering improvement suggestions aligned to organizational coding standards.

• Comprehensive PR Analysis: Analyzes diffs to find potential issues, anti-patterns, or deviations from best practices.
• Context-Aware Feedback: Offers targeted, actionable guidance based on the actual changes made in the pull request.
• Guideline Customization: Supports organization-specific review rules via editable guideline configurations.

Supported Input Mode:

• With Repository: Connect to a GitHub repository and choose the relevant branch for the pull request.

Additional Features:

• Edit Guidelines: Define or modify custom code review rules to tailor the agent's evaluation.
• Ask a Question Prompt: Use natural language to trigger a PR review by pasting the link into the question field.
• Diff-Level Intelligence: Feedback is anchored to exact lines and files changed in the pull request.

Example Questions You Can Ask:

• "Please review this PR for me: https://github.com/org/repo/pull/123"
• "Is there anything wrong with this pull request?"
• "Does this PR follow our error handling best practices?"
• "Does this PR include sufficient unit test coverage?"
• "Can you summarize what was changed in this PR?"
• "Suggest improvements for this pull request."

Automated code review is available through CI/CD pipelines using the Code Review Agent. To learn more. To learn more, see Automated Code Review in CI/CD.