OpenTag is a self-hosted AI agent that handles knowledge-work requests inside Slack and Microsoft Teams, using connected tools and generated interfaces. Teams can use it as an on-call triage bot or customize it for their own agent workflows. Its catalogue add-ons support the agent, tools, and integrations assembled in the starter application.
Borrowing it
Nothing to install: this file belongs to CopilotKit/OpenTag. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/CopilotKit/OpenTag/main/AGENTS.mdgit clone --depth 1 https://github.com/CopilotKit/OpenTagWrote 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/copilotkit/opentag/agents-md)<a href="https://agentmods.dev/instructions/copilotkit/opentag/agents-md"><img src="https://agentmods.dev/badge/instructions/copilotkit/opentag/agents-md/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/instructions/copilotkit/opentag/agents-md"><img src="https://agentmods.dev/badge/instructions/copilotkit/opentag/agents-md.svg" alt="Reviewed on agentmods" width="80" 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 5.1 | $0.02030 | $0.02030 |
| Opus 5.5 | $0.00812 | $0.00812 |
| Sonnet 5.5 | $0.00406 | $0.00406 |
| Haiku 4.5 | $0.00203 | $0.00203 |
Grade A, and why
OpenTag AGENTS.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 1mo 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 — 137 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Agent instructions
OpenTag is meant to be cloned and customized, often with a coding agent driving. This file tells that agent where the authority lives, so it grounds work in the current API instead of in recalled patterns.
Skills to install before you start
This repository vendors no skills. It installs them, because the CopilotKit CLI delivers them from one home and a vendored copy silently falls behind.
| Task | Skill | Install |
|---|---|---|
Changing Channel code — createChannel, handlers, commands, JSX message UI, modals, human-in-the-loop, runtime wiring |
copilotkit-channels |
npx copilotkit@latest skills install --skill copilotkit-channels |
Getting Slack to answer for the first time, or diagnosing a Channel stuck at setup_required or a mention with no reply |
setup-slack-channel |
npx copilotkit@latest skills install --skill setup-slack-channel |
| The whole setup path, including Microsoft Teams | channels-setup |
npx copilotkit@latest channels setup |
Runtime internals — CopilotRuntime, agent runners, tools |
runtime |
npx copilotkit@latest skills install --skill runtime |
setup-slack-channel is written for this repository. Its phases assume
OpenTag's conventions — app/channel.tsx, app/env.ts,
INTELLIGENCE_CHANNEL_NAME, a local agent on port 8123 — so prefer it here over
a generic setup sequence.
Always relay these with @latest. A bare copilotkit resolves to whatever is on
PATH or already in the npx cache, and an older CLI fails with
Unknown option '--skill'.
Do not commit what skills install writes. .agents/, .claude/skills,
agent/skills, and skills-lock.json are gitignored deliberately: those paths
are install targets, so committing them means the next install overwrites tracked
files.
Repository map
| Area | Path | What it owns |
|---|---|---|
| Runtime entrypoint | server.ts |
Environment, Channels readiness, HTTP lifecycle, shutdown |
| Composition | app/index.ts |
Agent factory, managed Channel, runtime |
| Channel surface | app/channel.tsx |
Mentions, commands, components, modals, interrupts |
| Environment contract | app/env.ts |
Required variables and in-code defaults |
| Rendered UI | app/components/, app/tools/ |
Issue cards, tables, charts, diagrams |
| Agent | agent/agent.py |
LangGraph deep agent served over AG-UI |
| AG-UI adapter | agent/agui.py |
Slack recursion limit and user-facing graph-stop handling |
| Persona | agent/prompts/ |
system.py is the base system prompt |
| Approval gate | agent/write_confirmation.py |
Emits confirm_write before a Linear, Notion, Composio, or Arcade write |
| Provider selection | agent/connected_app_provider.py |
Picks Composio or Arcade from which API key is set; both set fails startup |
| Composio | agent/composio_tools/ |
Toolkit sessions, per-person identity, effect classification, connect links. state.py also owns the trusted-actor rules for both providers |
| Arcade | agent/arcade_tools/ |
Catalogue, per-action authorization, effect classification, outcome handling, connect tickets and browser verification |
| Arcade connect pages | app/arcade-browser-routes.ts |
The two public pages of the Arcade connect flow. They live on the runtime so the agent keeps no public entry point |
| Coder | agent/coding/ |
GitHub credentials, Daytona sandbox, repository publish tools, coder prompt |
| Coder skills | agent/coding/skills/ |
Committed skills. Do not put them in agent/skills/ |
| Deployment | .railway/railway.ts |
Two services, declared as code |
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.
- 1mo ago Changed · +10 lines · +240 tokens per session e77c29a18baa
- 1mo ago Changed · +20 lines · +351 tokens per session 0502a98fea0a
- 2mo ago First seen · 107 lines · 1,439 tokens per session scan A 312832b54f5e
OpenTag AGENTS.md is an instructions file published in the GitHub repository CopilotKit/OpenTag (1,302 stars, last pushed 2d ago), licensed MIT. It adds 2,030 tokens to every session, about $0.0081 per session on Opus 5.5. 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.
OpenTag AGENTS.md compared with instructions from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| OpenTag AGENTS.mdCopilotKit/OpenTag | Instructions file | 1.3k | 2d ago | MIT | A | 2,030 | Codex, OpenCode |
| vscode buildNext.instructions.mdmicrosoft/vscode | Instructions file | 195k | yesterday | MIT | A | 9,719 | GitHub Copilot |
| next.js AGENTS.mdvercel/next.js | Instructions file | 143k | today | MIT | A | 7,461 | Codex, OpenCode |
| langchain AGENTS.mdlangchain-ai/langchain | Instructions file | 149k | yesterday | MIT | A | 4,680 | Codex, OpenCode |
| vscode oss-third-party-notices.instructions.mdmicrosoft/vscode | Instructions file | 195k | yesterday | MIT | A | 5,516 | GitHub Copilot |
| vscode copilot-instructions.mdmicrosoft/vscode | Instructions file | 195k | yesterday | MIT | A | 3,093 | GitHub Copilot |
| deepseek-harness AGENTS.mddeepseek-ai/deepseek-harness | Instructions file | 274k | 2d ago | MIT | A | 4,313 | Codex, OpenCode |