Reflexion
技能 已验证 活跃Reflect on previus response and output, based on Self-refinement framework for iterative improvement with complexity triage and verification
To significantly improve the quality, correctness, and completeness of LLM-generated outputs by embedding a critical self-evaluation and refinement loop.
功能
- Structured self-reflection process
- Complexity-based reflection depth triage
- Iterative refinement of LLM outputs
- Guidance on identifying and fixing deficiencies
- Automated reflection hooks based on prompt keywords
使用场景
- Improving code generation quality
- Refining documentation and explanations
- Ensuring completeness and correctness of LLM responses
- Applying rigorous quality gates to AI-generated work
非目标
- Directly executing code or external tools
- Replacing human oversight entirely
- Providing creative writing or novel content generation
工作流
- Assess task complexity (Quick, Standard, Deep Reflection Path).
- Perform Initial Assessment: Completeness, Quality, Correctness, Dependency & Impact, Fact-Checking, Generated Artifact Verification.
- Determine if Refinement is Needed.
- If needed, identify specific issues and propose prioritized solutions.
- If code is involved, check for existing libraries/solutions before custom implementation.
- Review architecture, code quality, and testing.
- Verify factual claims and assertions.
- Finalize output after self-reflection checklist and questions.
Documentation
- info:Configuration & parameter referenceThe `argument-hint` is mentioned but not fully documented; it suggests an optional focus area or confidence threshold but lacks specific examples or a clear default.
安装
请先添加 Marketplace
/plugin marketplace add NeoLabHQ/context-engineering-kit/plugin install reflexion@context-engineering-kit质量评分
已验证类似扩展
Migrate Validate
100Validate pending migrations for foreign key consistency, rollback safety, and best practices
Semgrep Rule Creator
100Creates custom Semgrep rules for detecting security vulnerabilities, bug patterns, and code patterns. Use when writing Semgrep rules or building custom static analysis detections.
Moyu (摸鱼)
100감지된 과잉 엔지니어링 패턴: (1) 사용자가 명시적으로 요청하지 않은 코드나 파일을 수정할 때 (2) 요청되지 않은 새로운 추상화 계층(클래스, 인터페이스, 팩토리, 래퍼)을 생성할 때 (3) 요청되지 않은 주석, 문서, JSDoc, 타입 주석을 추가할 때 (4) 요청되지 않은 새로운 종속성을 도입할 때 (5) 최소 편집 대신 파일 전체를 다시 작성할 때 (6) diff 범위가 사용자의 요청을 명백히 초과할 때 (7) 사용자가 "너무 많아", "거기는 건드리지 마", "X만 변경해", "간단하게", "그만"과 같은 신호를 보낼 때 (8) 발생할 수 없는 시나리오에 대한 오류 처리, 유효성 검사, 방어적 코드를 추가할 때 (9) 요청되지 않은 테스트, 설정 스캐폴딩, 문서를 생성할 때
Wrap Up Ritual
100End-of-session ritual that audits changes, runs quality checks, captures learnings, and produces a session summary. Use when saying "wrap up", "done for the day", "finish coding", or ending a coding session.
Cleanup Cycles
100Detect and untangle circular dependencies. Runs madge/skott (TS), pycycle (Py), or compiler-only checks (Go/Rust). Auto-fixes leaf-extractable cycles; reports core cycles for human review. Use when the user asks to find circular imports, fix dependency cycles, or untangle module graph. Example queries — "find circular imports", "fix dependency cycles", "untangle our module graph", "why is madge complaining".
Safe Mode
100Prevent destructive operations using Claude Code hooks. Three modes — cautious (warn on dangerous commands), lockdown (restrict edits to one directory), and clear (remove restrictions). Uses PreToolUse matchers for Bash, Edit, and Write.