When would you use an agent skill instead of a prompt, retrieval, or a tool?

Instruction: Distinguish reusable procedures, standing instructions, factual retrieval, and executable capabilities.

Context: Practice choosing where agent knowledge and procedures belong, including progressive loading and permissions.

Updated

Example Answer

I'd use a skill for a reusable procedure that needs instructions and supporting resources, such as preparing a report in a reviewed format. A prompt carries the task and standing rules, retrieval supplies relevant facts, and a tool performs an operation. They can work together.

I'd expose a short skill description first, then load its instructions and resources when needed. That avoids putting every procedure into every request. I'd version and evaluate the skill like other application behavior. Loading instructions doesn't grant permission to run a bundled script or send data; those actions still need the application's controls.

Concrete example

A reporting skill can describe the workflow and template. Retrieval finds the permitted sales records, and a tool creates the file. The skill supplies the procedure; it does not make restricted records accessible.

Follow-up to practice

What would make you replace a skill with a deterministic workflow?

Reference

Agent Skills: format and progressive disclosure

Your preparation path

Work through these questions in order. Read the answer aloud, then explain it in your own words.

1. Start with the foundations

Build the vocabulary and explain the core decisions.

  1. When would you use a deterministic workflow instead of an agent? Free sample
  2. What state should persist between agent turns? Free sample
  3. How do you decide whether an agent should ask a human for approval? Free sample

2. Apply it to a real workflow

Practice diagnosis, validation, and everyday tradeoffs.

  1. How would you keep an agent from looping on the same failed action? Free sample
  2. When do multi-agent systems create more risk than value? Member answer
  3. An agent keeps choosing the wrong skill. How would you diagnose and fix its catalog? Member answer

3. Prepare for senior discussions

Explain failure boundaries, recovery, and production choices.

  1. Design an agent that can resume after partial failure without repeating side effects. Member answer
  2. After a model upgrade, how would you decide which agent-harness rules to remove? Member answer
  3. An agent's execution container dies halfway through a task. What survives, and how does it recover? Member answer

Related Questions