MuseOn

Museon CLI gives the AI agent you already use the tools to research social content, understand what works, create new material, publish with your approval, and learn from real performance.

You keep working with your Agent. Museon gives it the social-media capabilities needed to move from an idea to real work.

The bundled base Skill defines Mel, your AI social-media operator. For Museon work, it guides your Agent to own the goal, execute through Museon CLI, verify results, and use real performance to improve the next round. This positioning is included in the CLI package and installed by museoncli setup.

Give this to your Agent

Copy the instruction below into Codex, Claude Code, Cursor, or another Agent that can install Skills and run shell commands:

Set up Museon CLI for this agent by following this onboarding guide exactly:
https://www.museon.ai/cli/install.md

That is the recommended installation path. You do not need to clone the repository, choose CLI flags, or configure credentials by hand.

What happens next

1. Your Agent installs Museon CLI. It installs the exact Python wheel from the official GitHub release and verifies that the command is available.
2. Museon CLI installs the Skills. The bundled, same-version workflow Skills teach the Agent how to route and use Museon safely and add the Instagram Hook research workflow. ego lite is installed separately when authenticated browsing is needed.
3. You approve access in the browser. Sign in to Museon and choose the workspace the Agent may use.
4. You describe the work, not the commands. The Agent discovers the right Museon actions and keeps the task moving.

Once connected, try a request like this:

Research the AI note-taking content gaining momentum on TikTok and Instagram.
Explain the repeated hooks and audience questions, then propose three carousel
ideas for our product. Do not publish anything until I approve it.

Agent capabilities

This working tree prepares the CLI 2.0 surface; the pinned installation below remains at the last published release until a separate release is approved.

Outcome	Domains
Research public content and Hooks	research
Plan, inspect, edit and verify HireAICreator content	hireaicreator
Track creators, posts and performance	campaign-monitor
Edit UGC briefs, brief assignments and which brief a creative cites	ugc
Analyze video content	content-analysis
Upload files, import image URLs and read media records	media
Validate, host and share durable reports	artifacts
Connect and inspect accounts, use bound phones, read performance and edit profiles	social-account
Schedule Agent work and retain run memory	routines
Discover workspace methodology	skills

Eight bundled Skills teach discovery, research, durable artifacts, account access, Hook analysis, monitoring and routines. The CLI executes atomic operations; the Agent composes workflows. Museon checks user, workspace membership, role and resource access for every request. Verify writes with readback from the owning domain, not with a request receipt alone.

See the migration guide before replacing an older CLI. The HireAICreator domain is included in this unpublished candidate.

You stay in control

• Research and other read-only work can run as part of the task.
• Creating, changing, scheduling, publishing, or deleting something requires the approval rules described by the command.
• The Agent must explain the intended change before a sensitive operation.
• Credentials stay in the Agent's local environment. Museon CLI uses the operating-system credential store when available and a mode-0600 file only in headless environments; an Agents-hosted session instead uses an exclusive, short-lived private lease that cannot fall back to a user API key. Authorization decisions remain on Museon's servers. See Authentication credential providers.

Install without an Agent

If you prefer to install the CLI yourself, use Python 3.11+, uv, and the exact wheel from the official GitHub release:

uv tool install "https://github.com/Museon-AI/museon-cli/releases/download/v0.22.2/museoncli-0.22.2-py3-none-any.whl"

Then continue with Skill setup and browser authorization:

museoncli setup --agent codex
museoncli auth start
museoncli auth finish --wait
museoncli skills +list
museoncli whoami

Use --agent claude-code or --agent cursor for those hosts; --agent auto uses the active host marker, or a single existing supported Agent home. If several Agent homes exist, choose one explicitly or use --agent all. Restart the Agent after installing the Skill. After authentication, skills +list shows the Business Skills available in the current workspace. auth finish --wait waits for approval for up to five minutes by default; use --timeout to change that limit. The shorter museon command is an alias for museoncli.

uv sync --frozen --all-groups
uv run ruff check .
uv run pytest -q
uv run python scripts/gen_command_docs.py --check
uv run python scripts/gen_command_contract.py --check
uv build --wheel
uv run python scripts/verify_public_artifacts.py

Command definitions live in museoncli/domains/. When a command changes, regenerate the documentation and portable command contract:

uv run python scripts/gen_command_docs.py
uv run python scripts/gen_command_contract.py

See CONTRIBUTING.md and the Code of Conduct to contribute. Release notes live in CHANGELOG.md. Please report security issues privately by following SECURITY.md.

License

Museon CLI is licensed under the Apache License 2.0.