agentleFS
GetStreamGetStream/agent-skills/skills/stream/SKILL.md

stream

Skill21 starsChanged 7 days agoApache-2.0View on GitHub

Stream overview, CLI, and router for Chat, Video, Feeds, Moderation, and Agents. Use for any Stream task: building or integrating Stream into web, iOS, Android, React Native, Flutter, Unity, or Unreal apps, building voice or text AI agents with Stream Agents, querying Stream data, running getstream CLI commands, looking up Stream SDK docs, or migrating (from Sendbird, or Feeds v2 to v3). Routes to the right skill.

  • Pipes a download into a shell

3 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

3 rules. 0 are common, 2 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 3 edited versions we read keep these.

  • 2

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Migrate Feeds v2 to v3: use stream-feeds-migration skill.

    line 80
  2. Distinctive:

    Migrate from Sendbird: use the platform-specific skill to migrate the code.

    line 81
  3. Very specific:

    Docs / "how does X work" questions: use the getstream docs command.

    line 84

“Common” is measured: kept by at least 6 in 10 of the 3 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Stream

Stream (getstream.io) provides APIs and SDKs for adding Chat, Video (calls and livestreaming), Activity Feeds, and Moderation to apps. A Stream account has organizations; an organization has apps; an app has one or more API key + secret pairs. All data - users, channels, calls, feeds - is scoped to an app. Client SDKs cover React, iOS, Android, React Native, Flutter, Unity, and Unreal; server-side SDKs mint user tokens and call the same API. Most client SDKs provide UI kits.

Stream Agents runs voice and text AI agents on Stream's backend. An agent is a folder of plain files (agents/<name>/agent.yaml, instructions, guardrail, knowledge, skills, simulations) synced with getstream agents, and sessions with it are opened through the Agents SDKs (@stream-io/calypso on the web). A session's conversation is a Chat channel and its voice is a Video call, so the app around an agent is a Chat and Video app like any other.

The CLI

The getstream CLI is the base tool for everything account- and data-side: authenticating, onboarding a project (org, app, credentials), calling any API endpoint, obtaining credentials, minting user tokens, reading the docs, and installing the other Stream skills. Don't work from memory of it - run getstream -h and getstream <command> -h and follow what they print.

To install (unless already present):

curl -fsSL https://getstream.io/cli.sh | bash
getstream --version

Alternatively, the user can install the CLI from getstream.io/cli.

The getstream agents commands are only in beta builds, whose version is a date and a commit (YYYY-MM-DD-<commit>) rather than semver. A stable build answers them with "agents unavailable". Install the beta over it:

curl -fsSL https://getstream.io/cli/beta/install.sh | bash

Routing

First match wins:

  1. Work on an app for a specific platform (build, integrate, audit, upgrade): use the platform-specific skill.

    • stream-react (React / Next.js)
    • stream-react-native (React Native / Expo)
    • stream-swift (iOS / SwiftUI / UIKit)
    • stream-android (Kotlin / Compose)
    • stream-flutter
    • stream-unity (Chat and Video)
    • stream-unreal (Chat only)

    For a review or audit of an existing integration ("is this ready for production?", "prepare for the integration review"), read integration-review.md together with the platform skill.

    Install a missing skill with getstream skills <name>, then follow it. For a web app, or a generic app, pick stream-react. A framework beats an OS: "an Unreal chat app for iOS" is stream-unreal.

    An app built around Stream Agents is platform work too, even when the request only says "agent" or "voice agent". Configure the agent from its folder and getstream docs agents/<sdk> (javascript, ios, android, flutter), and build the app with the platform skill: its Chat UI kit renders the session's conversation and its Video SDK holds the call. React Native, Unity, and Unreal have no Agents client SDK - tell the user before building.

  2. Migrate Feeds v2 to v3: use stream-feeds-migration skill.
  3. Migrate from Sendbird: use the platform-specific skill to migrate the code; to move the data itself (users, channels, message history), read sendbird-data-migration.md.
  4. Docs / "how does X work" questions: use the getstream docs command. In a command substitution like $(getstream docs <id>) don't add 2>&1. The path you need is on stdout.
  5. Query or mutate app data ("list channels", "any flagged messages?", one-off API calls): use the getstream api command.
  6. Genuinely ambiguous: ask.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.