untether-architecture

untether-architecture is a skill for Claude Code from littlebearapps/untether. It costs 51 tokens per session (2,204 once invoked), scanned A, original, MIT.

A reference explaining Untether’s architecture, the Telegram bridge that sends requests to coding agents and returns their progress and results.

In plain words
What is it for?
Use it when changing core infrastructure, registering an agent backend, following data flow, or understanding shared runner and event abstractions.
Why use it?
It shows how messages, runners, events, progress tracking, approvals, configuration, and cleanup fit together.

Skill for Claude Code

Written for Claude Code: installed under .claude/. Also seen: mentions Claude Code; mentions Codex; mentions OpenCode.

Not installable: its command points at a path on the author’s own machine, so it runs nowhere else. The line is /home/nathan/untether.

Install

Getting it into your agent

There is no command for this one: it runs only inside a plugin, and the catalogue could not identify which plugin ships it. The source is linked below.

Made for: Claude Code.

Wrote 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.

agentmods badge for untether-architecture

README.md
[![agentmods](https://agentmods.dev/badge/skills/littlebearapps/untether/untether-architecture.svg)](https://agentmods.dev/skills/littlebearapps/untether/untether-architecture)
Your own site
<a href="https://agentmods.dev/skills/littlebearapps/untether/untether-architecture"><img src="https://agentmods.dev/badge/skills/littlebearapps/untether/untether-architecture.svg" alt="Measured on agentmods" height="20"></a>
Per session 51 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 2,204 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5.1 $0.00051 $0.02204
Opus 5 $0.00026 $0.01102
Sonnet 5 $0.00010 $0.00441
Haiku 4.5 $0.00005 $0.00220

Measured 6d ago against content hash 24e9bda4cb0a, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-06, from the pricing page.

Security

Grade A, and why

untether-architecture 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 6d 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.

.claude/skills/untether-architecture/SKILL.md · 306 lines

How it starts

The opening of the file, as written. The whole thing — 306 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Untether Architecture

Telegram bridge for agent CLIs (Claude Code, Codex, OpenCode, Pi). Control coding agents from anywhere.

Data flow

Telegram Bot API
    |
    v
TelegramClient (httpx, long polling)
    |
    v
telegram/loop.py (parse updates, dispatch commands/callbacks)
    |
    v
handle_message() in runner_bridge.py
    |
    v
Runner.run(prompt, resume) -> AsyncIterator[UntetherEvent]
    |                              |
    v                              v
ProgressEdits <-- on_event() -- StartedEvent / ActionEvent / CompletedEvent
    |
    v
TelegramPresenter.render_progress() / render_final()
    |
    v
TelegramOutbox (coalesced edits, rate-limited sends)
    |
    v
Telegram Bot API

Core abstractions

Runner (Protocol)

class Runner(Protocol):
    engine: str
    def run(self, prompt: str, resume: ResumeToken | None) -> AsyncIterator[UntetherEvent]
    def is_resume_line(self, line: str) -> bool
    def format_resume(self, token: ResumeToken) -> str
    def extract_resume(self, text: str | None) -> ResumeToken | None

UntetherEvent (discriminated union)

type UntetherEvent = StartedEvent | ActionEvent | CompletedEvent

Every run emits: StartedEvent (once) -> ActionEvents (zero+) -> CompletedEvent (once, always last).

RunnerBridge (runner_bridge.py)

Connects runners to the transport layer:

  1. handle_message() — entry point for incoming Telegram messages
  2. Creates ProgressTracker and ProgressEdits
  3. Spawns runner in a task group with cancel support
  4. Manages progress message lifecycle (create -> edit -> replace with final)
  5. Handles errors, cancellation, ephemeral cleanup

ProgressTracker (progress.py)

Aggregates events into renderable state:

tracker = ProgressTracker(engine="claude")
tracker.note_event(evt)           # returns True if state changed
state = tracker.snapshot(
    resume_formatter=runner.format_resume,
    context_line="myproject@main",
)

Read the full file on GitHub · 306 lines

Changes

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.

  1. 6d ago First seen · 306 lines · 51 tokens per session scan A 24e9bda4cb0a

Subscribe to this mod's changes

untether-architecture is a skill published in the GitHub repository littlebearapps/untether (66 stars, last pushed 5d ago), licensed MIT. It adds 51 tokens to every session and 2,204 once invoked, about $0.0003 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.