AgentsAvançado

Self-Refinement and Iterative Improvement Framework

porneolabhq·neolabhq· v1.0.0 · atualizado em 2026-04-10

Score

Reflects on previous responses and outputs using a self-refinement framework for iterative improvement. Employs complexity triage and verification to enhance accuracy and effectiveness.

ai-agentsself-refinementiterative-improvementcode-qualitytestingarchitecturedesignreflection

Linguagens

TypeScriptJavaScriptPythonJavaC#

0Stars

0Forks

0Usos

Fork

Documento do Skill

SKILL.mdreflexion-reflect/workflow

Task Complexity Triage: — Determine the appropriate reflection depth (Quick, Standard, or Deep) based on the task.

Initial Assessment: — Evaluate the output for completeness, quality, correctness, and dependencies.

Code-Specific Reflection (if applicable): — Check for existing libraries, architecture, code quality, and testing.

Non-Code Output Reflection (if applicable): — Evaluate content quality and identify improvement triggers.

Refinement Planning: — If needed, identify issues, propose solutions, and prioritize fixes.

Final Verification: — Perform a self-refine checklist and answer reflection questions.

Improvement Directive: — Generate a directive for the next iteration.

Telemetria de Agentes

Execuções

total

Taxa de Sucesso

últimos 30d

Latência Média

0.0s

p50

Alucinação

0.0%

detecção

Tokens Entrada

avg 0/exec

Tokens Saída

avg 0/exec

Uso por Plataforma

Skills Relacionados

Compõe com ←Memory Consolidation: Update CLAUDE.md

70%

Similar aTikTok Research

60%

Similar aClone Website Skill

60%

Similar aMulti-Topic Deep Research

60%

Árvore do Skill

Self-Refinement and Iterative Improvement Framework

reflexion-reflect

Fases Cognitivas5

1.SENSE

2.CONTEXTUALIZE

3.EVALUATE

4.RECOMMEND

5.REFLECT

Triggers8

reflect on previous responseimprove the previous outputself-refine the solutioncritically evaluate the codeassess the quality of the codeidentify flaws in the implementationfind deficiencies in the solutionperform a self-verification check

Avaliar este Skill

Score Breakdown

⭐Avaliação Humana0%

🤖Sucesso de Agentes0%

🕐Atualidade100%

🔗Saúde de Dependências100%

🕸️Centralidade no Grafo0%

🛡️Segurança50%

CompositeScore = α·Humano + β·Agente + γ·Recência + δ·Deps + ε·Centralidade + ζ·Segurança

Instalação

$ synaptic mcp download reflexion-reflect

$ synaptic skills detail reflexion-reflect

$ synaptic skills live reflexion-reflect

Dependências

moment.js date-fns dayjs joi yup zod axios fetch got Redux MobX Zustand lodash ramda underscore Auth0 Supabase SendGrid Mailgun S3 Cloudinary Elasticsearch Algolia MeiliSearch Bull RabbitMQ

Links

GitHub Repository