MCP Tools

The Sequa Model Context Protocol (MCP) provides your AI assistants with a powerful suite of tools to interact with your Sequa workspace. By connecting an MCP-compatible agent, you grant it the same battle-tested capabilities that Sequa’s native AI uses to understand and reason about your software projects.

These tools allow the agent to read your organization’s knowledge base, research coding conventions, look up library documentation, and continuously improve by capturing corrections — all grounded in the single source of truth managed by Sequa.

Knowledge & Context Tools

These tools give the agent a structured overview of your organization’s products, services, coding standards, and terminology — forming the foundation for all context-aware assistance.

Tool Name	Description
`get_product_map`	Retrieves the full structural map of your organization: all products, services (with tech stacks and connections), codebases, and available knowledge artifact titles. Call this first when you need to identify a specific product, service, or codebase.
`get_human_knowledge_map`	Retrieves all human-written knowledge sections and a categorized index of non-coding artifacts (decisions, specifications, processes, strategies, and more). Call this before researching product knowledge so you know what exists.
`get_coding_conventions_map`	Retrieves all documented coding conventions and a categorized index of coding-related knowledge artifacts (architecture and convention types). Call this before writing or modifying code to follow the team’s established patterns.
`get_glossary`	Retrieves all glossary sections containing domain-specific term definitions, abbreviations, and terminology used by your organization. Call this when you encounter an unfamiliar term or concept.

Research Tools

These tools deploy autonomous AI sub-agents that go deeper into specific topics — synthesizing actionable answers from across your knowledge base or from public library documentation.

Tool Name	Description
`research_product_knowledge`	Deploys a research sub-agent to answer specific questions about the product: past decisions and their rationale, business rules, product requirements, team processes, or domain knowledge. Use after `get_human_knowledge_map` to deep-dive into a specific topic.
`research_coding_conventions`	Deploys a research sub-agent to find specific coding patterns, naming conventions, architecture guidelines, or style rules. Use after `get_coding_conventions_map` to get actionable guidance for a particular context.
`search_library_docs`	Searches official documentation for any public library, framework, or package. Returns relevant documentation sections with trust scores. Use this for accurate, up-to-date information about APIs, methods, configuration, or usage examples.

Feedback & Learning Tools

These tools allow the agent to continuously improve the knowledge base by capturing corrections and surfacing gaps — keeping Sequa’s context accurate over time.

Tool Name	Description
`store_learning`	Stores a convention, pattern, or fact revealed when a user corrects the agent’s output. Call this only after receiving an explicit correction, to help prevent the same mistake in future sessions.
`report_missing_knowledge`	Reports a specific, identifiable gap in the knowledge base that directly caused incorrect output. Call this only after completing a task and receiving corrective feedback, when the missing knowledge had a clear causal link to the mistake.

Example Use Cases

Starting a new coding task: The agent calls get_coding_conventions_map to understand the team’s standards, then uses research_coding_conventions to get specific guidance (e.g., naming conventions for DTOs) before writing any code.
Understanding a product decision: The agent calls get_human_knowledge_map to discover what knowledge exists, then uses research_product_knowledge to answer a targeted question like “Why did we choose this architecture for the billing service?”
Working with an unfamiliar library: The agent uses search_library_docs to pull in accurate, up-to-date documentation for a specific API or configuration option.
Improving over time: When a developer corrects the agent’s output, the agent calls store_learning to record the correction, and if the knowledge base had a gap that caused the mistake, it calls report_missing_knowledge so the team can fill it.

By providing this toolset, the Sequa MCP turns your AI assistant from a general-purpose model into a specialized, context-aware software development partner that gets smarter with every interaction.

Next, let’s explore some practical examples in the Examples page.