沟通五部曲
AdvisoryAudited by VirusTotal on Apr 17, 2026.
Overview
Type: OpenClaw Skill Name: five-steps-to-wisdom Version: 3.1.0 The skill bundle defines a structured communication and reasoning framework ('Five Steps to Wisdom') for an AI agent, focusing on vector retrieval, logical calibration, and knowledge management via Obsidian and ChromaDB. There is no evidence of malicious intent, data exfiltration, or unauthorized execution; the instructions in SKILL.md are purely behavioral guidelines for the agent's cognitive process.
Findings (0)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
When invoked, the agent may produce every answer in the five-step format rather than responding directly.
The skill imposes a mandatory response structure whenever it is used. This is aligned with the communication-method purpose, but it can override a user's expectation for a short or direct answer.
每次收到消息，回复前必须满足以下结构 ... ⚠️ 缺少任意模块 → 补上再发
Use this skill only when you want the five-step communication process, and explicitly ask the agent to shorten or skip the structure when it is not needed.
If the agent has memory, wiki, or web tools, conversation details and work results could be searched, reused, or saved for future tasks.
The instructions call for using vector memory, a wiki, web retrieval, and automatic storage into a memory system. This is consistent with the knowledge-loop purpose, but the artifact does not specify exact storage locations, retention, or approval boundaries.
三层检索顺序（向量 → wiki → Web Fetch） ... 五步法的产出自动沉淀 ... 存入记忆宫
Confirm which memory/wiki sources the agent can access, avoid sensitive content unless appropriate, and require confirmation before storing or reusing important private information.