Agent Prompt Enhancer FAQs

Question 1

Why does the skill flag unrestricted Bash tools?

Accepted Answer

Unrestricted Bash access violates the Principle of Least Privilege. The skill suggests scoped access, such as Bash(git:*) or Bash(npm:*), to prevent agents from executing unintended system commands.

Question 2

Which platforms are compatible with this skill?

Accepted Answer

The skill is designed to work with agent file structures for Claude Code (~/.claude/agents/), OpenCode (~/.config/opencode/agents/), and Codex (AGENTS.md).

Question 3

Can this skill automatically fix my prompt issues?

Accepted Answer

Yes, by using the --fix flag, the skill can automatically correct high-certainty issues such as missing frontmatter fields, unrestricted bash commands, and weak constraint language.

Question 4

How does it improve instruction effectiveness?

Accepted Answer

It identifies 'weak' language like 'should' or 'try to' and suggests stronger alternatives like 'MUST' or 'ALWAYS', while also encouraging positive instructions over negative ones.

Question 5

What is the 'Intern Test' for agent descriptions?

Accepted Answer

It is a validation check to ensure an agent's description clearly defines its triggers and exclusions, making it simple for a human or another AI to know exactly when to invoke it.

Agent Prompt Enhancer

主要功能

使用场景

Agent Prompt Enhancer

主要功能

使用场景