Albert: A model-agnostic AI coding CLI with provider fallback
Albert is a model-agnostic AI coding CLI tool built in Rust that operates as a terminal-native agent with a focus on security and autonomous decision-making. It uses a ternary logic system to validate actions before execution, ensuring safe and reliable interactions with the system. The tool supports multiple LLM providers, offers session persistence, and includes a suite of slash commands for development tasks while emphasizing sovereignty and context-aware cognition.
- ▪Albert is a Rust-built, terminal-native AI agent that uses a research-strategy-execute loop validated by ternary logic (+1, 0, -1) to ensure safe operations.
- ▪It supports multiple LLM providers including OpenAI, Anthropic, Google, and local models via Ollama, with no vendor lock-in.
- ▪Security features include deny-first AST command interception, zero-trust key storage, and protection against prompt injection and unauthorized command execution.
- ▪Albert includes persistent memory through a knowledge graph, session storage, and a user-editable identity file (ALBERT.md) that controls behavior and priorities.
- ▪The tool offers a comprehensive slash command library for tasks like planning, testing, code review, bug hunting, and pull request creation, with token optimization via RTK.
Opening excerpt (first ~120 words) tap to expand
Albert CLI The sovereign AI development CLI for the Ternary Intelligence Stack Albert is a terminal-native, model-agnostic AI agent built in pure Rust. It runs a hardened agentic loop — research → strategy → execute — with every action validated through ternary logic before anything touches your system. Quick Install — One Copy Gets It All # Install Albert (brings the full agent engine with it) cargo install albert-cli # Launch albert-cli That's it. Albert spins up the REPL and the onboarding will start.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at GitHub.