Package ai.mindconnect.adminui.ui.component
package ai.mindconnect.adminui.ui.component
-
ClassDescriptionTabbed detail view for a single agent: Details / Tools / Sessions.Edit form for an agent definition.Top-level list of agents: a search field, per-row Copy / Delete actions, header "New Agent" action, and pagination that carries the search filter.The chat page's attached-files panel: file name, chunk count and a Remove action per attachment.The chat-input form at the bottom of a
ChatPage— a single textarea plus a context-sensitive Send / Stop button.Read-only detail view of a single LLM configuration with Edit and Delete actions in the header.Edit form for an LLM configuration.Top-level list of all stored LLM configurations.Body of the "Test LLM config" modal dialog.Detail pane of the working-memory inspector — two tabs: Raw — the underlying record JSON via the json-viewer. LLM View — what the model actually sees: compressed stub if compression has run on the message, otherwise the original content.Master list of the working-memory inspector: every entry in the upcoming LLM call window — the system prompt plus all windowed messages — each as a clickable row that triggers a detail-pane replace via a GET on the?seq=…endpoint.The main conversation list of a chat page.The tool name and pretty-printed arguments out of a request's content JSON.Supplies the pre-built sub-agent cards (with their nested child-task trees) for a given parenttoolCallId.Lists pending migrations (bundled initial-data that is new or differs from what is stored), grouped into one collapsible section per entity type — the same layout the tool catalog uses.One LLM-call roundtrip rendered as a three-tab card: Response — prose first, then per-tool-call args (json-viewer) followed by the matching persisted tool-result block (collapsible).The "Info" tab of a chat page — a read-only key/value list of session metadata (id, agent, status, user, started-at, title) plus a "Back to Agent" link.Tabular view of an agent's chat sessions with Open / Delete row actions and a header "New Session" action.One tool-call or sub-agent-call card rendered as a collapsible<details>list item.The agent's checklist for the current session — same items the LLM sees in its prompt context via{{ todo_list_md }}.Read-only catalog of every tool the runtime'sToolRegistryknows about — the union of built-inToolFactorytools andMultiToolProvider-contributed tools.One catalog row: group (rubric), name, description, parameters schema and declared overrides schema.Body of the "Test tool" modal for the top-level tool catalog — the same shape asToolTestComponent, but identified by tool name (no owning agent).Read-only detail view of one agent's tool.Edit form for an agent's tool.Tabular view of an agent's tools.Body of the "Test tool" modal dialog.Detail-pane of the trace inspector — a stack ofRoundtripCardComponents for the selected turn, with consecutive sub-agent roundtrips wrapped in a collapsible<details>group so the parent stream stays readable.Master-list pane of the trace inspector — one row per top-level chat turn.One file listing for a single workspace scope (SESSION / AGENT_USER / USER).