Skip to content

Marketplaces & plugins

This is where agentsync earns its keep. A plugin is a bag of components — MCP servers, skills, subagents, commands, hooks, LSP servers. Install it once and every enabled agent gets the components it understands.

  1. Add a marketplace (a registry of plugins):

    Terminal window
    agentsync marketplace add github:anthropics/claude-plugins-official
  2. Install a plugin from it:

    Terminal window
    agentsync plugin install atlassian@anthropic
  3. Fetch from the networkupdate is the only networked verb:

    Terminal window
    agentsync update
  4. Apply — renders from the local cache to all agents:

    Terminal window
    agentsync apply

Each component in a plugin is projected independently per agent — fully, lossily, or skipped — and the report tells you exactly which:

plugin: atlassian@anthropic
claude ✓ full (1 mcp, 5 commands)
opencode ◐ partial (1 mcp; 5 commands → projected)

Inspect a plugin’s coverage without applying:

Terminal window
agentsync explain atlassian@anthropic
agentsync explain atlassian@anthropic --json

Limit a plugin to specific agents with an agents = [...] allowlist in its TOML file:

~/.agentsync/plugins/atlassian.toml
id = "atlassian@anthropic"
agents = ["claude"] # only fan this plugin out to Claude
Terminal window
agentsync plugin install <id[@marketplace]>
agentsync plugin upgrade <id>
agentsync plugin enable <id>
agentsync plugin disable <id>
agentsync plugin remove <id>
agentsync plugin list