npx agentspec-cli install 1070233b-fd89-425c-a534-68341814ab36 --file-type cursorrules→ .cursor/rules/
# CLAUDE.md — Cortex Agent Guide ## What is this? Cortex is an embedded graph memory engine for AI agents. Rust, redb (embedded DB), gRPC + HTTP APIs, vector search (HNSW), auto-linking. ## Repo structure ``` crates/ cortex-core/ — Storage, graph, vector, auto-linker, types. The library. cortex-server/ — gRPC/HTTP server, CLI, config. The binary. cortex-client/ — Rust gRPC client crate. cortex-proto/ — Protobuf definitions + generated code. warren-adapter/ — Optional Warren integration (feature-gated). sdks/ typescript/ — @cortex/client npm package python/ — cortex-memory PyPI package go/ — Go client module specs/ — Design specs (tagged IMPLEMENTED/DEFERRED) docs/ — User-facing documentation examples/ — Integration examples (LangChain, CrewAI, etc.) ``` ## Build & test ```bash cargo test --workspace # All 176 tests cargo test --workspace --no-default-features # Without warren adapter cargo build -p cortex-server --release # Release binary cargo clippy --workspace # Lint ``` ## Run ```bash cortex init # Create cortex.toml cortex serve # Start server (gRPC :9090, HTTP :9091) cortex node create --kind fact --title "..." cortex search "query" cortex briefing <agent-id> cortex shell # Interactive REPL ``` ## Key types - `NodeKind` / `Relation` — validated string newtypes (not enums) - `Cortex::open(path)` — library mode, no server needed - `Storage` trait — `RedbStorage` implements it - `IngestAdapter` trait — pluggable event sources - Config: `cortex.toml` with `#[serde(default)]` on all structs ## Architecture decisions - **redb** for storage (embedded, ACID, zero-copy mmap) - **HNSW** for vector search (instant-distance crate) - **gRPC** (tonic) for production API, **HTTP** (axum) for debug/viz - **warren-adapter** is optional (`--features warren`), cortex-core has zero network deps - Auto-linker runs background loop: similarity rules → edges, decay → prune, dedup → merge ## Config See `cortex.example.toml`. All config structs support partial overrides (serde defaults). ## Feature flags - `warren` (default) — enables Warren NATS integration - Without it: standalone graph memory with no network dependencies ## Specs Check `specs/` — all implemented specs are tagged. Deferred work tracked in GitHub issues. ## Tests - `crates/cortex-core/` — 128 unit tests (storage, graph, types, filters) - `crates/cortex-server/` — integration tests - `sdks/typescript/` — 28 Jest tests - `sdks/python/` — 30 pytest tests - `sdks/go/` — 9 Go tests ## Common pitfalls - Port 9090 may conflict with existing services — override in `cortex.toml` - Partial `cortex.toml` is fine — unset fields use defaults - `cortex-core` must have zero network dependencies — put network stuff in `cortex-server` or adapters - Node kinds and relations are lowercase strings with hyphens/underscores only
CLAUDE.md
This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
by AgentHub Crawler
AGENTS.md
Guide for AI agents working in this repository. **See also:** [DEVELOPMENT.md](DEVELOPMENT.md) - Development best practices and coding standards for AI-assisted development.
by AgentHub Crawler
AGENTS.md
Operational guide for AI agents working in this repository.
by AgentHub Crawler
AGENTS.md — agents-md-generator
`agents-md-generator` is a Python-based MCP (Model Context Protocol) server designed to generate and maintain `AGENTS.md` files—architectural context documents for AI coding agents. It uses Tree-sitter for high-fidelity AST analysis across multiple languages (Python, TypeScript, Go, C#) and `uv` for dependency management. The system operates as an incremental, cache-aware pipeline that transforms a codebase into a structured payload for AI assistants to perform maintenance tasks without manual e
Sign in to leave a comment.
by AgentHub Crawler