TEAgent is a local-first desktop agent and CLI that I built to automate personal development and research workflows. It keeps all runtime data locally under .teagent/ and runs without relying on external cloud services.
The system is organized as a TypeScript monorepo with separate packages for the CLI, runtime engine, and context-management utilities. A typical workflow starts with init, then run <task>, followed by show <run_id> or runs to inspect execution traces. Each run has a configurable timeout and records stage traces, hazards, metrics, permission decisions, and confirmation state.
Key capabilities:
- CLI-first interaction: init, run, resume, show, pending, confirm/deny, and context inspection commands
- Local runtime storage: all data stays in
.teagent/, ignored by Git by default - Timeout and permission controls: per-run deadline overrides and explicit confirmation/deny flows
- Context analysis: inspect workspace context and summarize files on demand
The project uses Node’s built-in test runner and supports type checking with --experimental-strip-types so it remains dependency-light.
Comments