Intent
Maintain continuity when no new data is present.
Context
Used when a user is chatting without uploading a new file and current-session features are empty. The system fetches recent historical insights from storage and injects them into chat context to avoid an “empty state” experience.
Agentic profile • System shape: unknown • Orchestration mode: unknown
Agent-to-agent interaction • Present: unknown • Mechanism: unknown • Evidence: unknown
Tool protocols • MCP: unknown • Tool calling: unknown • Evidence: unknown
Optimisation target • Primary: unknown • Secondary: unknown • Notes: unknown
Simplicity vs autonomy • Position: unknown • Rationale: unknown
Forces • Empty state handling • User experience
Solution • Detect when current session features/state are empty. • Fetch recent historical insights from persistent storage. • Inject the fetched items into the chat context for the next LLM call.
Implementation signals • Conditional logic checking for empty features • Retrieval from persistent storage
Evidence • src/agents/orchestrator.py#ADKOrchestrator.chat — Fetches recent insights if no upload • src/agents/chat_agent.py#ChatAgent._format_auto_insights — Formats historical items
Consequences
Benefits • Seamless user experience • Access to past context
Costs • Complexity in context management • Potential for stale data confusion
Failure modes • Injecting irrelevant history
Reuse notes • Check for empty session state and backfill with history. • Prefer retrieving only recent and relevant items to reduce stale-context risk.
Confidence
High — both the orchestrator behavior and the formatting helper are directly evidenced.