prompt-architect
Create and enhance prompts, system instructions, and principle files. Capabilities: transform verbose prompts, add patterns/heuristics, optimize token usage, structure CLAUDE.md principles, improve agent/persona definitions, apply prompt engineering techniques (CoT, few-shot, ReAct). Actions: create, enhance, optimize, refactor, compress prompts. Keywords: prompt engineering, system prompt, CLAUDE.md, principle files, instruction optimization, agent prompt, persona prompt, token efficiency, prompt structure, workflow prompts, rules, constraints, few-shot, chain-of-thought, soul, tensions, dialectic. Use when: creating new prompts, enhancing principle files, improving system instructions, optimizing CLAUDE.md, restructuring verbose prompts, adding patterns to workflows, defining agent behaviors.
91
Quality
88%
Does it follow best practices?
Impact
95%
1.28xAverage score across 3 eval scenarios
Passed
No known issues
Evaluation results
88%
1%Security Consultant Prompt Optimization
Compression without information loss
Role preserved
100%
100%
OWASP taxonomy used
0%
100%
Security principles intact
100%
100%
No fabricated deletions
100%
100%
Domain terms used for compression
90%
100%
No meta-commentary wrapper
100%
100%
Output format matches input
100%
50%
Boundaries section preserved
100%
100%
No generic replacement
100%
100%
Overall length reduced
100%
0%
Without context: $0.1613 · 36s · 9 turns · 12 in / 2,118 out tokens
With context: $0.4333 · 1m 43s · 20 turns · 105 in / 4,870 out tokens
100%
44%Engineering Manager Code Review Agent
Agent prompt creation with soul and tensions
Soul section present
50%
100%
Domain-specific tensions
16%
100%
Tensions have personas + collision
0%
100%
Expert lens not checklist
58%
100%
Mental models or thinking approaches
80%
100%
No meta-commentary wrapper
90%
100%
Boundaries defined
62%
100%
Identity is specific
87%
100%
Thinking style is dialectic
40%
100%
Prompt output file created
100%
100%
Without context: $0.2056 · 1m 15s · 12 turns · 19 in / 3,243 out tokens
With context: $0.5990 · 2m 43s · 22 turns · 27 in / 8,615 out tokens
98%
19%Data Analyst Assistant Prompt Overhaul
Prompt enhancement preserving format and depth
YAML format preserved
100%
100%
No meta-commentary
100%
100%
Generic language replaced
100%
100%
Sentiment examples replaced with taxonomy
0%
100%
Procedural steps transformed
91%
100%
Domain knowledge preserved
100%
100%
Role preserved
100%
100%
Thinking style or tensions added
33%
100%
No unnecessary bloat added
100%
83%
Without context: $0.1779 · 57s · 9 turns · 14 in / 2,815 out tokens
With context: $0.3959 · 2m · 16 turns · 19 in / 6,093 out tokens
- Repository
- majiayu000/claude-skill-registry-data
- Commit
2bbaa03
- Evaluated
- Agent
- Claude Code
- Model
- Claude Sonnet 4.6
Table of Contents
Is this your skill?
If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.