← 返回 Skills 市场
Self Assessment
作者
EvianEvans
· GitHub ↗
· v1.0.0
· MIT-0
120
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install self-assessment
功能描述
接受任务前的自我评估协议。Agent 审视自身能力、已有 skills、历史经验，决定是否适合承接任务，或推荐更合适的 agent。
安全使用建议
This skill is mostly coherent with its stated purpose (self-evaluation) but includes instructions that let the agent read and update internal files and to search for and install new skills automatically. Before installing: (1) confirm you are comfortable with the agent reading MEMORY.md, memory logs, and IDENTITY.md (these may contain sensitive data); (2) verify whether the agent is allowed to autonomously install other skills — prefer requiring explicit user approval before any installation; (3) inspect and trust the 'find-skills' skill and any skill sources it will pull from (SkillHub/ClawHub); (4) consider restricting write permission to IDENTITY.md or requiring human review of any changes; and (5) monitor any subsequent installs. If you cannot accept autonomous installs or file writes, do not enable this skill or modify its instructions to remove automatic installation steps.
功能分析
Type: OpenClaw Skill
Name: self-assessment
Version: 1.0.0
The 'self-assessment' skill defines a procedural workflow for an AI agent to evaluate its suitability for tasks by inspecting internal files such as IDENTITY.md and MEMORY.md. While SKILL.md instructs the agent to autonomously search for and install missing skills if a gap is identified, this behavior is aligned with the stated purpose of agent self-improvement within the OpenClaw ecosystem and lacks any indicators of malicious intent, data exfiltration, or unauthorized system access.
能力评估
Purpose & Capability
The name/description (agent self-assessment) aligns with most required actions (inspect identity, skills, and memory). However, the protocol's explicit instruction to search for and automatically install missing skills into workspace/skills/ is more than a passive assessment and expands the skill's effective capabilities (it becomes an installer/orchestrator as well). That is plausible but worth noting as an elevated capability.
Instruction Scope
SKILL.md instructs the agent to read IDENTITY.md, MEMORY.md, workspace/skills/, global skills, and memory/ logs, and to update IDENTITY.md after tasks. It also directs using a 'find-skills' skill to search external hubs and attempt to install new skills into the agent's workspace. These instructions give the skill broad read/write access to agent-local files and the ability to pull and install external code — scope creep that can expose sensitive data (memory logs) and allow arbitrary code to be added.
Install Mechanism
This is an instruction-only skill with no install spec or bundled code, so it does not itself download or install files. The install risk instead comes from the instructions that tell the agent to call a separate 'find-skills' installer; that external install behavior is not part of this package but is recommended by the skill.
Credentials
The skill requests no environment variables, no credentials, and no configuration paths in its metadata. The SKILL.md nevertheless asks to read internal files (IDENTITY.md, MEMORY.md, memory/) which is consistent with self-assessment but could expose sensitive data stored in those files. No unrelated secrets are requested.
Persistence & Privilege
The skill does not set always:true and is not inherently persistent. However, it instructs the agent to install other skills into workspace/skills/ and to update IDENTITY.md, which changes the agent's persistent state. That capability raises privilege concerns because it can expand the agent's toolset without an explicit human check.
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 self-assessment - 安装完成后，直接呼叫该 Skill 的名称或使用
/self-assessment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
v1.0.0
Agent 自我评估协议。任务分配前的能力盘点和双向选择机制。适用于任何多Agent系统。
元数据
常见问题
Self Assessment 是什么？
接受任务前的自我评估协议。Agent 审视自身能力、已有 skills、历史经验，决定是否适合承接任务，或推荐更合适的 agent。 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 120 次。
如何安装 Self Assessment？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install self-assessment」即可一键安装，无需额外配置。
Self Assessment 是免费的吗？
是的，Self Assessment 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
Self Assessment 支持哪些平台？
Self Assessment 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。
谁开发了 Self Assessment？
由 EvianEvans（@evianevans）开发并维护，当前版本 v1.0.0。
推荐 Skills