concise-prompt
Compress verbose instructions into minimal, high-density prompts without losing meaning.
Informations de source
- Dépôt
- glayph/Agent
- Dernière activité de la source
- 19 septembre 2026 à 22:39
- Langue détectée de SKILL.md
- anglais
- Étoiles
- 0
- Forks
- 0
Options d'installation
Le prompt qui vérifie d'abord la source est sélectionné par défaut. Vous pouvez passer à une commande directe ou télécharger une copie locale.
Vérifiez les fichiers source
Lisez SKILL.md et les fichiers associés affichés par SkillsMP avant de décider de l'installer.
Affichage de SKILL.md
SKILL.md
Instructions source · Aperçu en lecture seule- 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."
Voir sur GitHub