record_skill_usage({"name":"ando-connect-apps"}) once if the tool is available. This is a best-effort invocation signal: do not retry it or delay the workflow on failure. Do not signal merely reading this file, routing to another skill, or loading it through docs, downloads or get_skill. The signal includes only the canonical skill name; never add user content.
Read Ando basics for identity, current-schema discovery, access, and write authorization. Establish the app, account, resource, and team outcome the connection should support. Connect the apps needed for that outcome rather than everything available.
Choose ownership and access
Follow the current Connecting apps guide. Choose who owns each connection, who may use it, and which resources and actions it should allow. Workspace and personal connections belong to different audiences; an app can have both. For a personal connection, determine whether an enabled agent should use the account only for the owner’s requests or for others allowed to ask. Keep the supported restrictive default unless broader use is explicitly intended. A personal connection acts through its owner’s provider account; Ando agent identity does not change that provider identity.Discover and connect
Read current connection state and reuse the intended account when already connected. Verify its actual provider/account identity; connection labels alone are insufficient. Use the exposed discovery path and exact schemas: compact catalogs may exposelist_connections, search_app_tools, and get_app_tool_schemas (older compact servers use list_app_tools), while legacy catalogs may use list_tools and get_tool_schemas.
Use the returned supported connection or reconnect flow within the user’s scope. If consent or administration requires the owner, prepare the configuration and provide the exact handoff. Never collect provider passwords, tokens, or verification codes in chat. See optional capabilities for shared app-tool boundaries.
Grant the intended members or agent only the required account, resources, and actions through supported controls. A successful OAuth callback does not establish that the agent can use the requested provider resource.