Reflexion

Reasoning Schema

Before analysis: feature name, stage, iteration number, feedback items, previous patterns.

After analysis: root causes identified, reflections stored, patterns checked, retry guidance generated.

Invariant Principles

1. Every Failure Teaches: ITERATE verdicts contain actionable information.
2. Patterns Over Instances: Single failures are learning; repeated failures are patterns.
3. Root Cause Focus: Symptoms are feedback; causes are lessons.
4. Knowledge Accumulates: Reflections persist across iterations and features.
5. Guidance Prevents Repetition: Next attempt must address previous failure.

Inputs / Outputs

Phase Sequence

Steps 1-3: Full Analysis Pipeline

Dispatch subagent with command: reflexion-analyze

The subagent executes the complete analysis pipeline:

1. Parse Feedback - Extract structured fields from each feedback item
2. Categorize Root Cause - Map failures to root cause categories (Incomplete Analysis, Misunderstanding, Technical Gap, Scope Creep, Quality Shortcut, Integration Blind Spot)
3. Root Cause Questions - Answer expected vs actual, why deviation occurred, what would have prevented it
4. Store Reflections - Write to forged.db with PENDING status
5. Generate Retry Guidance - Produce specific correction guidance for the re-invoked skill

Pattern Detection

Integration with Develop Workflow

Trigger: forge_iteration_return with ITERATE verdict

Flow: Roundtable ITERATE -> forge_iteration_return -> reflexion skill -> analyze + store + check patterns + generate guidance -> return to develop orchestrator -> re-select and re-invoke skill

Escalation

After 3 iterations on same stage with same root cause: mark ESCALATED, report attempts made, recommend human intervention.

Example

1. Parse: source=Hermit, severity=blocking, stage=IMPLEMENT
2. Categorize: Quality Shortcut (missing validation)
3. Root cause: Rushed implementation, skipped security checklist
4. Store reflection with status=PENDING
5. Pattern check: Hermit flagged validation 2x before -> alert
6. Generate guidance: "Add input validation to all endpoints before resubmit"

Self-Check

• All feedback items analyzed for root cause
• Root causes categorized (not just described)
• Reflections stored with PENDING status
• Pattern check performed
• Retry guidance includes specific corrections
• Escalation evaluated if iteration >= 3

If ANY unchecked: complete before returning.

<FINAL_EMPHASIS> Failure is information. The roundtable said ITERATE because something was wrong. Your job is to understand WHY, not just WHAT. Store the lesson. Check for patterns. Guide the retry. The same mistake twice is repetition, not learning. </FINAL_EMPHASIS>