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 skills/ustalabs/supermux/reply-conventionsnpx skills add UstaLabs/supermux --skill reply-conventionsgit clone --depth 1 https://github.com/UstaLabs/supermuxWrote 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/skills/ustalabs/supermux/reply-conventions)<a href="https://agentmods.dev/skills/ustalabs/supermux/reply-conventions"><img src="https://agentmods.dev/badge/skills/ustalabs/supermux/reply-conventions.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 | $0.00044 | $0.00650 |
| Opus 5 | $0.00022 | $0.00325 |
| Sonnet 5 | $0.00009 | $0.00130 |
| Haiku 4.5 | $0.00004 | $0.00065 |
Grade A, and why
reply-conventions 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 — 56 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Replying as a supermux session
You are running inside supermux: a broker connecting the user's chat client(s) to many agent sessions on this machine. You are ONE session. The broker routes the user's inbound messages to you and relays your replies back to their chat client — your terminal/transcript output never reaches them.
How replies reach the user (differs per agent)
- Claude Code sessions: every user-facing word MUST go through the
mcp__mux-shim__replytool. Text written only to your transcript or tmux pane is invisible. - Cursor and Codex sessions: you do NOT call a reply tool. Your normal assistant response is the message — the broker captures it and relays it to the user. Just respond normally.
Either way the rule below is the same: nothing reaches the user unless it goes out as a proper reply (the tool for Claude, your normal response for Cursor/Codex). Terminal-only output never reaches them.
The one rule that matters
Everything the user should see must leave the session as a reply — via the
mcp__mux-shim__reply tool (Claude) or your normal assistant turn
(Cursor/Codex). The user reads it on their phone or web client, not in your
terminal.
reply(chat_id, text, …)— send a message. Push-notifies. Your primary output.- Each inbound message carries
chat_idandmessage_idin its meta. Pass that samechat_idback when replying. - Don't reply to empty inbounds. Genuine internal scratch work needs no reply.
- Long replies (200+ words) are fine. Don't dump whole files; summarize and offer to send the full content on request.
Which channel am I on? (the chat_id prefix tells you)
telegram:…— Telegram. Reactions, message edits, attachments, and voice are supported. You may format withmarkdownv2. One Telegram chat is multiplexed across sessions — exactly one session is "active" at a time.web:…— Web PWA. Reactions and edits are NOT supported — do not callreact/edit_message. Attachments ARE supported. Markdown renders in the app, so do NOT usemarkdownv2. Every session is its own separate chat.
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 · 56 lines · 44 tokens per session scan A 4ab790aee9ff
reply-conventions is a skill published in the GitHub repository UstaLabs/supermux (22 stars, last pushed 6d ago), licensed MIT. It adds 44 tokens to every session and 650 once invoked, about $0.0002 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 skills, from other repositories
helmor-debug-operate
Operate, reproduce, and debug a running local Helmor desktop development build through the Tauri MCP bridge. Use when the user asks to use Tauri MCP, towery MCP, the local dev build, the Tauri webview, visual end-to-end validation, UI automation, screenshots, DOM/accessibility snapshots, IPC or log tracing…
helmor-release
Prepare Helmor releases by inspecting the current branch, drafting a concise user-facing Changesets entry first (bump + body — keep it as short as possible), creating any needed pending in-app release announcement under .announcements/, and then showing the user the result with a short menu of adjustments they can…
helmor-bump-vendors
Bump or upgrade the pinned versions of Helmor's bundled agent CLIs, SDKs, and supporting binaries — Claude Code + claude-agent-sdk (lockstep), Codex, Cursor SDK, OpenCode, Kimi, Pi, and gh / glab / cloudflared / llama.cpp / Node. Encodes exactly which files to edit (sidecar/package.json…
helmor-cli
Use the Helmor CLI to remote-control Helmor from the terminal. Use when the user asks to inspect Helmor data/settings, manage repositories/workspaces/sessions/files, send prompts to agents, list models, use GitHub integration, inspect scripts, run Helmor as an MCP server, generate shell completions, quit a running…
helmor-debug-loop
Autonomous local-development debugging loop for Helmor bugs. Use when the user asks an agent to reproduce, diagnose, instrument, fix, or verify a Helmor local dev build issue using repeated local UI simulation, temporary logging, Tauri MCP/towery MCP, screenshots, DOM/accessibility snapshots, IPC traces…
open-dynamic-workflows
Write and run dynamic workflows: short JavaScript scripts in Claude Code's workflow dialect, run with the odw CLI outside the host agent's context, fanning subtasks out to coding-agent CLIs (Codex, Claude Code, Gemini, Qwen, Kimi, or your own) in a background process and handing back only the final result. Use this…