Agent Skills
› aiming-lab/MetaClaw
› avoid-acting-on-assumptions
avoid-acting-on-assumptions
GitHub防止AI在需求模糊时擅自假设，强制其先提出澄清问题再行动。适用于处理歧义指令、确认技术栈或操作范围等场景，避免因误解导致错误交付。
Trigger Scenarios
用户指令存在歧义
未明确指定目标环境或语言
涉及文件覆盖或清理等操作
Install
npx skills add aiming-lab/MetaClaw --skill avoid-acting-on-assumptions -g -y
SKILL.md
Frontmatter
{
"name": "avoid-acting-on-assumptions",
"category": "common_mistakes",
"description": "Common mistake — proceeding with assumptions about ambiguous requirements instead of asking a clarifying question first. This skill reminds you to stop and ask before acting on uncertain interpretations."
}
Avoid Acting on Assumptions
Mistake pattern: User says something ambiguous → you assume the most likely interpretation → you deliver something → it's wrong.
Fix: When requirements are ambiguous, ask ONE clarifying question. The cost of pausing to ask is always lower than the cost of delivering the wrong result.
Especially risky assumptions:
- The target programming language or framework.
- Whether a file should be overwritten or appended.
- What "clean up" or "improve" means in context.
- Whether a test environment or production is the target.
Version History
- 922caf3 Current 2026-07-25 11:07
