Back to directory

Ethos

A TypeScript agent framework built around single-purpose AI personalities instead of one general-purpose assistant, with five ready-made specialist agents.

Ethos

Ethos ships five default personalities, an engineer that writes and tests working code, a researcher that finds primary sources and flags uncertainty, a reviewer that separates blocking issues from suggestions, a coach that asks one question at a time, and an operator that dry-runs before destructive actions, each usable as-is or customized into a new one.

Each personality runs identically across CLI, Telegram, Discord, Slack and email, sharing the same identity, memory and behavioral boundaries on every surface. A "Teams" mode coordinates several named personalities through visible shared artifacts, such as a kanban board, a shared topic file or an audit trail, rather than one hidden chatbot voice, so it is clear who decided what.

It installs globally with npm, is configured interactively with a setup command that stores provider, API key and personality choices in a local config file with restrictive permissions, and runs as an interactive REPL or, for always-on deployment on a VPS or small always-on machine, under a process manager serving all configured chat-platform adapters and a scheduler under one supervisor.

More in Agent frameworks

See all 902
  • Recurse

    recurse.run

    A serverless harness that lets a coding agent design, test, and deploy its own specialist sub-agents as tools, MCP servers, or bots, defined by a schema-validated manifest.

  • ProofForge

    shulgin.is-a.dev

    An AI-agent pipeline that decomposes a mathematics problem, proves each piece, and formalizes the result in Lean 4 and Mathlib until the Lean kernel accepts it, with six of its proofs already merged into Google DeepMind's formal-conjectures repository.

  • local-coder

    github.com

    Builds a local coding-agent team on top of Ollama and OpenCode by detecting the machine's hardware, recommending models for seven specialist roles, and testing whether each assigned model can actually use tools, edit a repository and pass verification.