The Agent Skills Directory

[COMMAND_EXECUTION]: Skill guides the agent to perform configuration tasks including file writes to .cursor/mcp.json, .mcp.json, and .vscode/mcp.json to enable Model Context Protocol integrations.- [REMOTE_CODE_EXECUTION]: For environments where hosted MCP is unavailable (such as Federal or EU regions), the skill provides instructions to run @launchdarkly/mcp-server via npx. This utilizes the vendor's official package from the standard npm registry.- [EXTERNAL_DOWNLOADS]: Directs the agent to configure connections to LaunchDarkly's hosted MCP servers at mcp.launchdarkly.com and references official documentation from Cursor, Claude, Windsurf, and Microsoft.- [CREDENTIALS_UNSAFE]: References the use of LAUNCHDARKLY_ACCESS_TOKEN for local server authentication. The skill includes robust security guidance, advising users to add configuration files to .gitignore and implementing a mandatory consent step with a security warning before the agent assists with handling the secret.

mcp-configure