termul: Instructions file for Claude Code

AGENTS.md

Termul loads AGENTS.md from the workspace root as agent memory (similar to AGENTS.md / CLAUDE.md). This file is also the project's living architecture doc — read it before making changes.

Project

Termul (TERminal mULtiplexer) — open-source AI-native terminal & dev workspace. Tauri 2 + Rust (portable-pty) backend, React 19 + TypeScript + xterm.js (webgl) client, BYOK AI via Vercel AI SDK v6.

• Bundle id: app.binsarjr.termul
• Package manager: pnpm
• Platforms: macOS, Linux, Windows
• Frontend checks: pnpm exec tsc --noEmit, pnpm test
• Rust checks: cd src-tauri && cargo clippy && cargo test --locked

Quality bar

Production-grade or it does not ship. Every change is judged against all of these, not just "it works":

• Correctness: edge cases, failure modes, concurrent access. No "works for now".
• Performance: ultra-lightweight is the product. ~7-8 MB bundle, high-performance terminal. For every change ask: how much RAM it costs, whether it adds IPC round-trips or redundant requests, whether it triggers extra re-renders or wasted work, whether it pulls a heavy dependency. Unused features consume zero resources. Lightweight-first is non-negotiable: if a requirement or proposed change risks added latency or lag (input-to-render, scrolling, large/streamed output, extra re-renders or IPC round-trips, a heavy dependency), STOP and raise it with the user before implementing. Do not silently absorb the cost — speak up first and discuss the trade-off.
• Security: no critical security holes. Validate at every boundary (IPC, fs, network, AI tool surface). The secret-path deny-list applies on both read and write and is never bypassed.
• UI/UX: polished, professional, premium. Every state and detail considered.
• Architecture: new or changed logic lives in pure, dependency-light functions (functional core); tauri commands and React components stay thin (imperative shell). Keeps it testable without a later rewrite.