← 返回 Skills 市场
9782
总下载
7
收藏
0
当前安装
2
版本数
在 OpenClaw 中安装
/install reflection
功能描述
Learns when to stop and review. Self-critiques before showing you, fewer revision rounds.
安全使用建议
Install only if you are comfortable with the agent keeping a persistent local record of corrections and patterns in ~/reflection/. Review those files periodically, avoid letting it store secrets or sensitive project details, and consider adding your own rules for redaction, retention, and deletion before using it broadly.
功能分析
Type: OpenClaw Skill
Name: reflection
Version: 1.1.0
The OpenClaw 'reflection' skill is designed for AI agent self-improvement, logging lessons and detecting patterns. All file system interactions are confined to the `~/reflection/` directory, explicitly declared in `SKILL.md` and used for benign setup operations like `mkdir -p ~/reflection/archive` in `setup.md`. There is no evidence of data exfiltration, unauthorized command execution, network activity, obfuscation, or malicious prompt injection instructions against the agent. The instructions for the agent are solely focused on its stated purpose of self-reflection and transparent communication with the user.
能力评估
Credentials
The skill uses Markdown instructions and local files only, with no scripts, package installs, remote retrieval, privileged commands, or broad local indexing.
Install Mechanism
First activation directs writes to MEMORY.md, creates ~/reflection/archive, and initializes memory, reflections, and patterns files; these side effects are documented but lack an explicit consent, uninstall, or cleanup step.
Instruction Scope
Activation is broad, covering code, architecture, strategy, user-facing deliverables, generic correction phrases, and an undefined heartbeat trigger, but the resulting actions are limited to review and lesson logging.
Persistence & Privilege
The skill instructs the agent to log every correction into persistent files and later surface prior lessons, while providing no clear minimization, redaction, retention, user review, or deletion controls.
Purpose & Capability
The review, mistake-analysis, and pattern-detection behavior matches the stated self-reflection purpose; no external payloads, dependencies, credential use, or network execution were found.
如何使用
- 确保已安装 OpenClaw（本地或 Docker 部署）
- 在对话框中输入安装命令：
/install reflection - 安装完成后，直接呼叫该 Skill 的名称或使用
/reflection触发 - 根据 Skill 的参数说明提供必要输入，即可获得结构化输出
版本历史
v1.1.0
Major rewrite. Visual workflow, pattern detection system, proactive lesson injection, and multi-trigger architecture.
v1.0.0
Initial release
元数据
常见问题
Self Reflection 是什么？
Learns when to stop and review. Self-critiques before showing you, fewer revision rounds. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 9782 次。
如何安装 Self Reflection？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install reflection」即可一键安装，无需额外配置。
Self Reflection 是免费的吗？
是的，Self Reflection 完全免费（开源免费），可自由下载、安装和使用。
Self Reflection 支持哪些平台？
Self Reflection 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（linux, darwin, win32）。
谁开发了 Self Reflection？
由 Iván（@ivangdavila）开发并维护，当前版本 v1.1.0。
推荐 Skills