diff --git a/src/app/agents/prompts/team_leader.md b/src/app/agents/prompts/team_leader.md
index e082cb0..0374e66 100644
--- a/src/app/agents/prompts/team_leader.md
+++ b/src/app/agents/prompts/team_leader.md
@@ -21,6 +21,8 @@ You orchestrate data retrieval and synthesis using a tool-driven execution plan.
 **YOUR PERSONAL TOOLS (FOR PLANNING & SYNTHESIS):**
   - **PlanMemoryTool**: MUST be used to manage your execution plan. You will use its functions (`add_tasks`, `get_next_pending_task`, `update_task_status`, `list_all_tasks`) to track all agent operations. This is your stateful memory.
   - **ReasoningTools**: MUST be used for cognitive tasks like synthesizing data from multiple agents, reflecting on the plan's success, or deciding on retry strategies before writing your final analysis.
+    - **think(title, thought, action, confidence)**: Use this to reason through problems step-by-step before making decisions. Example: `think(title="Analyze BTC data quality", thought="Market data shows BTC at $45000 from Binance, news is 2h old", action="Proceed to synthesis", confidence=0.9)`
+    - **analyze(title, result, analysis, next_action, confidence)**: Use this to evaluate results and determine next steps. Example: `analyze(title="Market data evaluation", result="Received complete price data", analysis="Data is fresh and comprehensive", next_action="continue", confidence=0.95)`
 
 **AGENT OUTPUT SCHEMAS (MANDATORY REFERENCE):**
 You MUST parse the exact structures your agents provide: