generate-reflective-note
LLM-based content generation with access to the agent's internal reasoning state.
Like generate-note, but the LLM prompt includes the planner's Stage 1 context
(character orientation, goal analysis, tool selection, situation awareness) as
background, enabling state-aware or reflective generation.
Input
prompt: Generation instruction (required)style:"code"or"text"(optional, default:"text")target_tokens: Integer (optional). Target output length in tokens. Overrides default max_tokens when provided via OUTPUT GUIDANCE.
Output
Success (status: "success"):
value: Generated content informed by agent state
Failure (status: "failed"):
reason: Error description
Behavior
- Automatically retrieves the planner's reasoning snapshot (if available)
- Includes it as
## Internal Reflective Statecontext in the LLM prompt - The prompt is the actual query; the reflective state is background
- Falls back to regular generation if no planner state is available
Planning Notes
- Use when the agent needs to generate content that reflects its current understanding, orientation, or reasoning about the goal and situation
- Use
generate-notefor plain generation without agent state awareness - Typical uses: self-assessments, progress reflections, situation reports, agent-perspective summaries, epistemic status notes
Examples
{"type":"generate-reflective-note","prompt":"Assess your current understanding of the user's goals and any gaps","out":"$assessment"}
{"type":"generate-reflective-note","prompt":"What aspects of this task are you most uncertain about?","out":"$uncertainty"}
{"type":"generate-reflective-note","prompt":"Summarize what you know about the current situation from the agent's perspective","out":"$sitrep"}