"mcpServers": {
    "playwright": {
      "type": "stdio",
      "command": "npx",
      "timeout": 30,
      "args": [
        "-y",
        "@playwright/mcp@latest"
      ],
      "disabled": false
    }
  }
}```
So I’m not planning to change anything inside the existing Arguments field itself, but instead add a separate tab for editing/viewing the JSON configuration.

I also asked Claude to generate a rough design mockup. Does this align with the desired behavior or am I missing something? Thank you!

<@U0975SLALTT> <@U0975QYCNRF>

• add configurable chat context compaction (#4535) (dabdb34)
• add resume/reconnect chat feature (#4219) (f14f3d3), closes #3012
• *catalog:* add Sensitive column to MCP Headers table (#4840) (11ac36b)
• *catalog:* share cascade-reinstall decision between frontend and backend (#4841) (ea8e400)
• *chat:* allow bare skill slash commands (#4940) (05c130d)
• *chatops:* add system prefix with provider and thread id (#5053) (44378b2)
• *chatops:* enrich Slack LLM context with channel id, workspace id, and permalink (#5059) (ec6a413)
• *code-runtime:* Dagger-backed run_python tool + local dev (#5018) (db27a93)
• *code-runtime:* deploy managed Dagger runtime via kube-pod (#5019) (f6a2251)
• default theme to light (#4883) (3eaf14e)
• *dev-stack:* auto-copy .env from the main worktree (#5027) (cbccf5d)
• *dev-stack:* hydrate parallel stack's DB from main (#5044) (84f27e2)
• frontend pre-send sensitive data detection (#5024) (fd8e166)
• improve knowledge connector options (#5060) (1beb4f3)
• length-aware entropy threshold and URL skip in sensitive data detector (#5028) (f4d6f47)
• mcp apps sidebar (#5037) (d237afb)
• *oauth:* scope reverse-proxy host trust to getPublicRequestOrigin (#4965) (c460cf4)