concise-prompt
Compress verbose instructions into minimal, high-density prompts without losing meaning.
소스 정보
- 저장소
- glayph/Agent
- 최근 소스 활동
- 2026년 9월 19일 22:39
- 감지된 SKILL.md 언어
- 영어
- 스타
- 0
- 포크
- 0
설치 방법
기본적으로 소스를 먼저 확인하는 Prompt가 선택됩니다. 직접 명령으로 전환하거나 로컬 사본을 다운로드할 수도 있습니다.
소스 파일 검토
설치 여부를 결정하기 전에 SKILL.md와 SkillsMP에 표시된 보조 파일을 읽어 보세요.
SKILL.md 표시 중
SKILL.md
소스 지침 · 읽기 전용 미리보기- name
- concise-prompt
- description
- Compress verbose instructions into minimal, high-density prompts without losing meaning.
- version
- 1.0.0
- author
- Miki
- license
- MIT
- platforms
- ["linux","macos","windows"]
- metadata
- {"Miki":{"tags":["prompt-engineering","compression","token-efficiency"],"related_skills":["semantic-summarization"]}}
# Concise Prompt Engineering (Instruction Compression)
## Principles
- Remove all filler words, hedging, and redundancy
- Use terse, imperative structure
- Prefer single words over phrases
- Eliminate polite formalities ("please", "could you")
- Use abbreviations where unambiguous
- Flatten nested instructions into ordered lists
## Compression ratio target
- Aim for 60-80% reduction in token count
- Never sacrifice task correctness for brevity
## Technique patterns
1. **Ellipsis**: Remove subjects/objects implied by context
- "You should analyze the code" → "Analyze code"
2. **Imperative stacking**: Chain actions without connectors
- "First do X, then do Y" → "Do X. Do Y."
3. **Category labels**: Replace descriptions with known labels
- "fix syntax errors in TypeScript" → "tsc --noEmit"
4. **Token-efficient framing**: Use `{ }` templates for variable slots
5. **Delete meta-commentary**: Remove "I'll", "Let's", "We need to"
## Example
Input (78 tokens):
"You should carefully review the entire codebase and find any bugs that might exist, and then provide a detailed report of what you found."
Output (15 tokens):
"Review codebase. Find bugs. Report findings."
GitHub에서 보기