Anthropic has added a feature that lets Claude users connect plugins automatically when they log in, provided the plugin's MCP server accepts Claude as a trusted issuer.
Claude · Plugins & MCP
Plugins now connect automatically when you sign in to Claude
Supported plugins connect at login and stay connected without repeated authorization. The feature depends on each plugin's MCP server accepting Claude as a trusted issuer, and developers choose whether to add support.
How the connection now works
A simplified flow from sign-in to a connected plugin
Step 1
User signs in
Login to Claude is the only step the user takes.
→
Step 2
Plugin connects
The MCP server trusts Claude as the issuer, so no separate connection step is needed.
→
Step 3
Stays connected
No repeated re-authorization on each use.
What it offers
Faster start for end users with supported plugins
Fewer repeated authorization prompts
Less friction for developers who opt in
What is still unclear
Rollout timing
Which servers or plugins already support it
How administrators and enterprise customers manage these connections
Independent developer and security reviewer reaction
Key condition
Login-time connection works only where the MCP server accepts Claude as a trusted issuer.
Trust: the server honors Claude's identity assertions.
Opt-in: developers decide whether to add support.
Trade-off: convenience now, with login and permission design becoming more central as agents adopt external tools.
Continue reading The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in ✓ Signed in — this article isn’t included in your current plan.Unlocking the full article...