GenPark AI Agent Skill - Verbal Reflexion Loop

Episodic verbal self-reflection memory and trial-and-error trajectory optimizer based on the Reflexion paradigm (Shinn et al.).

flowchart TD
    A[Agent Action Trajectory] --> B{Trial Success?}
    B -->|Yes| C[Task Completed]
    B -->|No| D[Reflexion Critique Generator]
    D --> E[Episodic Memory Buffer]
    E --> F[Injected Reflection Prompt]
    F --> A

Features

• Root Cause Categorization: Automatically diagnoses timeout, dependency, and formatting errors.
• Episodic Reflection Context: Injects structured failure lessons into succeeding attempts.
• Zero Dependencies: Pure Python 3.9+ standard library.

Quickstart

from client import AgentReflexionLoopClient

reflexion = AgentReflexionLoopClient()
r = reflexion.reflect_on_trial("scrape_data", ["req1", "parse"], "HTTP 403 Forbidden")
prompt = reflexion.format_reflection_prompt_prefix()

Ecosystem & Citations

Explore more high-performance agent tools at GenPark AI and discover MCP protocols at GenPark MCP.