What it does
CodeCortex indexes a repository's structure, symbols, definitions, references, dependencies and call relationships, along with Git history, ownership and architecture. For a given task, it ranks evidence drawn from lexical, semantic, structural, graph, Git and memory signals, so the agent receives a focused package instead of broad file reads. It can list reverse dependencies, affected tests and change risk before an edit, and it offers semantic edits and structural rewrite previews with explicit review steps. Project memory, traces and multi-repo workspaces persist beyond a single chat.
It suits work where an agent needs to navigate an unfamiliar codebase, check what a change might affect before editing, or keep architectural decisions across sessions. The project includes configuration for Claude Code, Codex, Cursor, Gemini CLI and OpenCode. The configurator manages only its own MCP entries and leaves other user configuration unchanged.
Running it requires Python 3.11 to 3.13. After installation, the repository can be exposed to an MCP-capable coding agent. A bundled deterministic demo indexes a fixture project, analyses the blast radius of a sample class, routes an evidence request and reports measured context and trace data. The publisher records a hardening test suite of 711 passed, 28 skipped and 0 failed, with 91.74% coverage, and documents the scope and methodology in HARDENING_REPORT.md.
Limitations and requirements
Requires Python 3.11 to 3.13. The recorded hardening figures are committed measurements for a specific test scope and are not generalised performance promises. Scope and methodology are described in the publisher's HARDENING_REPORT.md.
Safety
Our team has not reviewed this resource. Only automated rules checked it; resources with high-risk signals are not published. These are signals, not a guarantee that installing it is safe. Try it in a sandbox project first.
- Runs code or changes an assistant's behaviour
Sources
- Official MCP Registry first seen 13 Nov 2026 · last seen 13 Nov 2026 · its page on the source
CodeCortex Context Engine
Context intelligence for AI coding agents: navigation, impact, memory, guarded edits.
v0.1.0a10As shown on registry.modelcontextprotocol.io when we last read it