instruction-engineering
Enhances prompt design for AI agents by applying research-backed techniques to improve accuracy and effectiveness.
Install this skill
or
96/100
Security score
The instruction-engineering skill was audited on Feb 12, 2026 and we found 4 security issues across 1 threat category. Review the findings below before installing.
Categories Tested
Security Issues
low line 15
External URL reference
SourceSKILL.md
| 15 | 1. **Emotional Stimuli Work**: [EmotionPrompt](https://arxiv.org/abs/2307.11760) (Microsoft, 2023): +8% instruction induction, +115% BIG-Bench. [NegativePrompt](https://www.ijcai.org/proceedings/2024/ |
low line 19
External URL reference
SourceSKILL.md
| 19 | 3. **Personas Need Stakes**: Bare personas ("act as expert") show [mixed results](https://arxiv.org/abs/2311.10054). Persona + emotional stimulus ("You are a Red Team Lead. Errors will cause serious p |
low line 49
External URL reference
SourceSKILL.md
| 49 | **Research:** [Large Language Models Understand and Can Be Enhanced by Emotional Stimuli](https://arxiv.org/abs/2307.11760) (Microsoft, William & Mary, 2023) |
low line 61
External URL reference
SourceSKILL.md
| 61 | **Research:** [NegativePrompt: Leveraging Psychology for Large Language Models Enhancement](https://www.ijcai.org/proceedings/2024/719) (IJCAI 2024) |
Scanned on Feb 12, 2026
View Security DashboardGitHub Stars 106
Rate this skill
Categorydevelopment
UpdatedMay 21, 2026
majiayu000/claude-skill-registry