Releases

Derive cbor2 serde impls for CBOR protocols: integer map keys, semantic tags and inspectable COSE metadata.

CLI harness for autoconf-rs tools: autoconf, autoheader, autom4te, autoreconf, aclocal, autoscan, autoupdate, ifnames

Core M4 macro engine and shell script generator for autoconf-rs (clean-room forensic-parity)

AgentTemplate — a small functional DSL executed on top of the Cersei runtime

Tool trait, built-in tools, and permission system for the Cersei SDK

Future Reth integration boundary for eth.

Sandbox & VM isolation for the Cersei coding-agent SDK

Future signer isolation boundary for eth.

Future RPC trust-policy boundary for eth.

Placeholder for trusted publishing setup

Future REVM adapter boundary for eth.

Core macro expansion engine for m4-rs: lexer, parser, evaluator, builtins.

Memory trait and backends for the Cersei SDK

agentskills.io-compatible skill registry for the Cersei SDK

Magi Agent: a terminal AI assistant in Rust with sandboxed tool execution, OAuth login, and encrypted local memory (Argon2 + AES-256-GCM-SIV + Reed-Solomon FEC).

Provider trait and built-in LLM providers for the Cersei SDK

Optional sanitization bridge for eth secret-bearing Ethereum data.

MCP (Model Context Protocol) client for the Cersei SDK

Hook/middleware system for the Cersei SDK

Optional derive macros for eth-valkyoth-sanitization.

A minimal and clean task manager for your terminal

An accurate natural language detection library, suitable for short text and mixed-language text

Sync engine for the saku ecosystem

Stack analysis for AVR

The Zulu language model for Lingua, an accurate natural language detection library

The Yoruba language model for Lingua, an accurate natural language detection library

The Xhosa language model for Lingua, an accurate natural language detection library

The Welsh language model for Lingua, an accurate natural language detection library

Derive macro for the Cersei Tool trait

The Ukrainian language model for Lingua, an accurate natural language detection library