codeatlas-mcp-server

Auto-generate a Mermaid flowchart diagram showing how modules, classes, and functions connect in the system. Returns a Mermaid diagram string that AI can read to understand the full system flow without reading every file.

Perform deep code analysis on a local project directory. Generates AST analysis in memory and syncs to CodeAtlas Cloud.

List all projects that have been analyzed by CodeAtlas. Returns project names, paths, and last analysis time.

Get all modules, classes, functions, and variables in the analyzed project. Returns entity type, name, file path, and line number.

Get import/call/containment/implements relationships between entities. Shows how modules, classes, and functions are connected.

Get AI-generated code insights including refactoring suggestions, security issues, and maintainability analysis.

Search for functions, classes, modules, or variables by name. Supports fuzzy matching.

Get all entities (classes, functions, variables) defined in a specific file.

Create or update the .agents/memory/ folder with auto-generated system documentation. This folder serves as AI

Retrieve the auto-generated system documentation and episodic memories (business rules and change logs) for a project from CodeAtlas Cloud / Oracle 26ai.

Save a dream memory (mistake, preference, knowledge, or pattern) to CodeAtlas Cloud for long-term AI recall. The AI uses this to persist learnings across conversations.

Query previously saved dream memories from CodeAtlas Cloud. Uses semantic search to find relevant memories based on the query text. Returns memories with relevance scores.

Check dream memory sync status. Returns count of stored dreams grouped by type and project. Use this to verify dreams are syncing to the cloud correctly. Can be called from any AI IDE, CLI, or Hermes cron.

Search CodeAtlas Genome for relevant genes. Uses semantic search to find the most relevant genes.

Get a specific gene by ID from the CodeAtlas Genome.

Scan the CodeAtlas Immune System for previously encountered failures matching a problem description. Returns prevention context to inject into prompts.

Record a failure pattern as an immune gene in CodeAtlas Genome. This helps prevent future agents from repeating the same mistake.

Trace the complete flow of a feature through the codebase. Given a keyword (e.g.

Generate a Mermaid diagram showing the EXECUTION FLOW of a feature. Unlike generate_system_flow (which shows module imports), this traces the actual call chain: entry point → controller → service → model → database. Given a keyword, it finds all related functions and classes, then builds a flowchart or sequence diagram showing how they call each other at runtime. This is the best tool for understanding HOW a feature works step-by-step.

Knowledge Graph Reasoning: Use Oracle 26ai Graph features to automatically detect architectural weaknesses, circular dependencies, God objects, and dead code.

Enterprise Scanner: Automatically scan all analyzed projects for bugs, security vulnerabilities (hardcoded secrets, unsafe functions), and architectural problems. Features Admin Insights and Security Scoring.

Search source FILE CONTENTS across the entire project for any text string. Unlike

Find ALL functions, methods, or classes that call or reference a specific symbol. The

Find everything a function, method, or class calls or depends on. The

Full BLAST RADIUS analysis for changing a symbol. Traces BOTH callers (what depends on this) AND callees (what this depends on) in one view. Also finds related test files. Use BEFORE any significant code change.

Get a comprehensive overview of a project in ONE call: package.json (name, version, scripts, deps, devDeps), config files detected, README summary, test framework, git branch. Saves 5-10 individual read_file calls when starting work.

Run an npm/pnpm/yarn script from package.json. Returns exit code, stdout/stderr, and duration. Handles cd to project dir automatically.

Get recent git changes: last N commits (hash, author, date, message, files changed), uncommitted changes (modified/added/deleted), branch status (ahead/behind). Saves multiple git commands.

Configure CodeAtlas Second Brain for any MCP client (Hermes, Claude Code, Gemini CLI).

Check the current Second Brain configuration status for all MCP clients.

Manage Architecture Decision Records (ADRs) — persistent decisions that survive context resets. Use when documenting architecture choices, listing past decisions, updating status, or checking what was decided and why. ADRs persist at ~/.codeatlas/adr/<project>/.

Read source code of a specific function/class/module by its qualified name. Returns the exact file path, line range, and raw source. Use when you need to see the actual implementation — not just that it exists.

Check what files and entity types are indexed for a project. Returns: file count, entity distribution, files with most entities, and files that might be missing (not indexed). Use after

Find near-duplicate or semantically similar functions/classes in a project. Uses token-based Jaccard similarity to find code that looks different but does the same thing. Returns groups of similar functions with similarity scores. Use before refactoring to consolidate duplicated logic.

Export a compressed snapshot of the project

Scan all installed skills (from ~/.agents/skills/ and ~/.claude/skills/) and save a complete inventory to CodeAtlas Second Brain. This lets the AI remember what skills are available, query them, and reference them by name. Run after installing new skills or on first setup.