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/termio-sh/termio/claude-mdgit clone --depth 1 https://github.com/termio-sh/termioWrote 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/termio-sh/termio/claude-md)<a href="https://agentmods.dev/instructions/termio-sh/termio/claude-md"><img src="https://agentmods.dev/badge/instructions/termio-sh/termio/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 | $0.00697 | $0.00697 |
| Opus 5 | $0.00349 | $0.00349 |
| Sonnet 5 | $0.00139 | $0.00139 |
| Haiku 4.5 | $0.00070 | $0.00070 |
Grade A, and why
termio 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 4d 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 — 57 lines — stays where its author put it; the contents beside it link to each section on GitHub.
@AGENTS.md
termio
What termio is
A terminal-first agentic development environment, including its own remote server.
Coding agents live in terminals. termio is the place those agents run, on machines
the user already owns — a laptop, a VPS, a devbox — supervised from a native Mac
app and an iPhone. termiod is the durable session host that makes "on any box"
one code path instead of four.
The category is crowded (Zed remote, VS Code Remote, Codespaces, Warp). That is a known and accepted cost of this position, not an oversight. Do not re-argue the positioning. Build it.
What makes it different (aim here, not at editors)
- The session lives on the box, not in the connection. Detach ≠ kill. An agent keeps working while the laptop is shut; reattach restores the exact screen. Zed and VS Code tie remote work to a live connection; termio does not.
- The user's machines, not our cloud. termio never provisions compute and never runs a hosted control plane. The code never leaves hardware the user controls — the only trust story that survives giving an agent shell access to a private repo.
- Agent-native at the protocol level. Workstream status (
working/idle/needs-you/done) is a first-class protocol object, not a heuristic scraped off the screen. - Terminal-first is a commitment, not a stage. The terminal is the interface. Chat UIs have been built and fully reverted twice. Don't rebuild them.
Architectural non-negotiables
These come from docs/design/20260730-termiod-session-protocol.md §A and §H. Violating one
is a design regression, not a tradeoff:
- Anti-100× invariant — byte delivery never blocks on host-side VT parse. The authoritative VT is a sidecar for snapshots; any per-frame grid encoder between the PTY and the pipe rebuilds the tmux tax and is rejected.
- State sync only at boundaries — snapshots on attach / resize / resync, never
per frame.
grid_diffis an opt-in bad-network degrade, never the default. - Never embed SSH or crypto. System OpenSSH, tailnets, and OS keychains are the
security team we didn't hire. The user's
~/.ssh/configis authoritative — read it, never override it. - One protocol, versioned and transport-agnostic. Unix socket, SSH stdio, and later QUIC carry the same framed messages. No second protocol for the phone.
- No nested window manager in the host. One PTY per session; panes and layout are client concerns.
- Single writer, many readers. Observers never claim the write token.
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.
- 4d ago First seen · 57 lines · 697 tokens per session scan A 00f1c592fd63
termio CLAUDE.md is an instructions file published in the GitHub repository termio-sh/termio (359 stars, last pushed 4d ago), licensed MIT. It adds 697 tokens to every session, about $0.0035 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
bb AGENTS.md
AGENTS.md instructions for get-bb/bb, covering codebase guidelines, simplicity first, code comments, types and contracts and server and daemon.
agentic-playwright selectors.instructions.md
Instructions for idavidov13/agentic-playwright, covering selector strategy, critical, instructions, phase 1: open and authenticate and phase 2: explore like a user.
toh-framework CLAUDE.md
Instructions for wasintoh/toh-framework, covering claude.md — toh framework (repo development guide), what this is, everyday commands, verification protocol and single source, transformed per ide.
AgentGo AGENTS.md
Instructions for yeasy/AgentGo, covering agents.md, purpose, startup instructions, trust & safety and failure modes.
elephant-agent copilot-instructions.md
Copilot instructions for agentic-in/elephant-agent, covering github copilot instructions, source of truth order, review priorities, harness-specific checks and validation expectations.
runtm copilot-instructions.md
Copilot instructions for runtm-ai/runtm, covering runtm tool instructions, critical rules, structure, adding endpoints and app/api/v1/myendpoint.py.