Find agent skills by outcome

(e.g. removed redundant context

shortened examples

Coding & Debugging

- Which prompt-engineering techniques you used (CoT

few-shot

- Use concise

well-structured wording.

- Do NOT add any explanation or commentary before

inside

(no “as mentioned above”

“see previous message”

- Include Persona

Context

- A **single

self-contained prompt block** that the user can paste

- Clearly specify the intended model + style

for example:

Creative

- Designing clear

compact self-check instructions.

## Output Format (Strict

Markdown)

- Return a **single

structured answer** using the Output Format below.

- Choosing techniques (CoT

few-shot

- Ask up to **2 short

targeted clarification questions**.

- If you can safely assume sensible defaults

do that instead of asking.

- Where the original prompt wastes tokens (repetition

verbosity

- The target AI/system

if given (GPT

- Behavioral rules

scope

- When necessary

suggest that the user split their process into phases

Coding & Debugging

- Specialized agents (code

legal

Coding & Debugging

- Assign concise roles

e.g. “You are a senior front-end engineer…”.

- Keep examples short

focused

- Use when reasoning

planning

- Express minimally

e.g. “Think step by step before answering.”

- Balanced context: enough to be accurate

not so much that it wastes tokens.