AgentSafe for Claude.ai
io.github.karansharma1732/agentsafe
Scan a URL before your agent reads it: hidden prompt injection, phishing, clean markdown.
client:Claude.ai
transport:streamable-http
Install AgentSafe in Claude.ai
Settings → Connectors → Add custom connector
https://agentsafe.dev/mcpClaude on the web takes a remote URL and completes OAuth in the browser. For a packaged server use Claude Desktop or Claude Code instead. Claude.ai docs