Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add instructions/flipbit03/terminal-use/claude-mdgit clone --depth 1 https://github.com/flipbit03/terminal-useWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/instructions/flipbit03/terminal-use/claude-md)<a href="https://agentmods.dev/instructions/flipbit03/terminal-use/claude-md"><img src="https://agentmods.dev/badge/instructions/flipbit03/terminal-use/claude-md.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00655 | $0.00655 |
| Opus 5 | $0.00328 | $0.00328 |
| Sonnet 5 | $0.00131 | $0.00131 |
| Haiku 4.5 | $0.00065 | $0.00065 |
Grade A, and why
terminal-use CLAUDE.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 5d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 43 lines — stays where its author put it; the contents beside it link to each section on GitHub.
terminal-use (tu)
Headless virtual terminal for AI agents. Binary is called tu. See issue #1 for the full project spec.
Build & Test
cargo build # dev build
cargo build --release
cargo test # run unit tests
cargo fmt --check # check formatting (CI enforces this)
cargo clippy -- -D warnings # lint (CI enforces this)
RUSTDOCFLAGS="-Dwarnings" cargo doc --all-features --no-deps # doc lint (CI enforces this)
Architecture
CLI client → Unix socket (JSON-over-newline) → background daemon → per-session PTY + alacritty_terminal emulator.
src/main.rs— clap CLI, command dispatchsrc/daemon/server.rs— Unix socket listener, daemon lifecycle.WaitandMouserequests run outside the manager lock so they don't blockmonitorpolls.src/daemon/session.rs— PTY + emulator per session, plus a per-sessionMouseTracker(synthetic cursor, held buttons, last event)src/daemon/manager.rs— session map, request handling,handle_mouse_glided(interpolated mouse motion)src/daemon/protocol.rs— JSON request/response types, mouse action/target/state typessrc/emu.rs— wrapper aroundalacritty_terminal+ its embeddedvte::ansiparser; exposes the small Parser/Screen/Cell/Color slice the rest of the codebase consumes. Capture proxy queuesEvent::PtyWritereplies (DA, DCS, etc.) for the reader to forward back to the PTY.src/pty/— PTY spawn (ECHOCTL off so curses subshell apps see clean ESC echo), input injection, resizesrc/keys.rs— key name → escape sequence mappingsrc/mouse.rs— wire encoders for SGR / Default / UTF-8 mouse protocols, screen-search helpers (find_text/find_regex)src/commands/— CLI command handlers (each talks to daemon)src/render/— text and image screenshot renderers (PNG paints a magenta△cursor overlay)
Conventions
- Follow lineark patterns:
usagecommand for LLM reference,version = "0.0.0"patched by CI,binary-releasefeature flag. usage≠--help.usageis a hand-maintained <1000 token cheatsheet for agents.--helpis verbose clap output for humans.- Commands that don't need the daemon (usage, self update) are handled before daemon connection.
- Default terminal size: 120x40, TERM=xterm-256color, COLORTERM=truecolor (overridable via
--env COLORTERM=...).
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 5d ago First seen · 43 lines · 655 tokens per session scan A c11db06f97ca
terminal-use CLAUDE.md is an instructions file published in the GitHub repository flipbit03/terminal-use (87 stars, last pushed 18d ago), licensed MIT. It adds 655 tokens to every session, about $0.0033 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
tmuxy CLAUDE.md
Claude Code instructions for flplima/tmuxy, covering tmuxy, project structure, cli usage, pane operations and tab operations.
terminal-jarvis AGENTS.md
AGENTS.md instructions for BA-CalderonMorales/terminal-jarvis, covering agents.md - terminal jarvis, current shape, key sections, branch strategy and merge-to-develop discipline.
mars AGENTS.md
Instructions for Yazelix/mars, covering mars agent guide, workflow, source boundaries, validation and release workflow.
jmux CLAUDE.md
Claude Code instructions for jarredkenny/jmux, covering claude.md, project, commands, architecture and the two-channel tmux model.
termul AGENTS.md
Instructions for binsarjr/termul, covering agents.md, project, quality bar, conventions and releases.
termul CLAUDE.md
Instructions for binsarjr/termul, a project described as: Termul (TERminal mULtiplexer) — a lightweight, AI-native terminal & dev workspace. Tauri 2 + Rust + React. No telemetry, no account.