Few-Shot Example
Demonstration inputs and outputs included in prompts to show Claude the expected format, tone, and quality level rather than describing them abstractly.
Few-shot examples are the most reliable way to calibrate Claude's output. Instead of describing the desired tone ("write like an operator"), you show 2-3 examples of operator writing. The /produce-content skill includes before/after examples of anti-slop corrections. The /persuasive-copywriting skill includes examples of PVP (Provoke, Validate, Propose) messaging. Few-shots work because they ground abstract instructions in concrete output patterns. The practical limit is 3-5 examples per skill: more than that wastes context tokens, fewer than 2 leaves too much room for interpretation. Examples should cover the common case and one edge case.
Where it shows up:
Related Terms
Methodology
Prompt Engineering
The systematic design of AI instructions for consistent, high-quality output, moving beyond ad-hoc p...
Skills Layer
Skill
A reusable instruction set that encodes a repeatable process. Skills load context, apply methodology...
Methodology
Anti-Slop
A 4-category quality framework that detects and eliminates AI-generated writing patterns: banned phr...
Architecture
Context Window
The token limit within which Claude operates per conversation, determining how much knowledge, histo...
Related Terms
Prompt Engineering
The systematic design of AI instructions for consistent, high-quality output, moving beyond ad-hoc prompting to structur...
TermSkill
A reusable instruction set that encodes a repeatable process. Skills load context, apply methodology, and produce consis...
TermAnti-Slop
A 4-category quality framework that detects and eliminates AI-generated writing patterns: banned phrases, structural sin...
TermContext Window
The token limit within which Claude operates per conversation, determining how much knowledge, history, and instructions...