Intent
Provide relevant state to the agent for context-aware responses.
Context
Used when a chat-facing agent must respond to user queries about their data, while LLM calls themselves are stateless. This pattern suits systems that keep state externally (stores, DB) and need to present it to the LLM at call time.
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 • Stateless LLM calls • Need for personalization
Solution • Gather relevant state from stores before the agent call. • Inject that state into the prompt template as structured text using prompt variables. • Keep the injected context aligned to what the agent needs to answer the user query.
Implementation signals • Prompt string formatting with variables • Data retrieval before agent call
Evidence • src/agents/chat_agent.py#ChatAgent.run — Injects features, insights, plan, and summary
Consequences
Benefits • Context-aware answers • No need for fine-tuning
Costs • Context window limits • Prompt complexity
Failure modes • Context truncation • Information overload
Reuse notes • Aggressively retrieve and format context before calling the LLM. • Keep injected context concise and structured to avoid prompt bloat.
Confidence
High — the chat agent run method is explicitly evidenced as injecting multiple context elements into the prompt.