We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/stevengonsalvez/promptregistry-mcp'
If you have feedback or need assistance with the MCP directory API, please join our Discord server
# β
Prompt Refactoring Complete
Successfully refactored all prompt files from embedded JSON content to clean markdown + metadata structure!
## π New Structure
```
default_prompts_data/
βββ agentic-coder-enhanced.md (15KB - Clean markdown content)
βββ agentic-coder-enhanced.json (1.5KB - Metadata only)
βββ memory-bank-assistant.md (4.7KB - Clean markdown content)
βββ memorybank-assistant.json (843B - Metadata only)
βββ nextjs-codebase-review.md (3.1KB - Clean markdown content)
βββ nextjs-codereview-assistant.json (681B - Metadata only)
βββ python-codebase-review.md (3.3KB - Clean markdown content)
βββ python-codereview-assistant.json (668B - Metadata only)
βββ prompt-writing-assistant.md (2.3KB - Clean markdown content)
βββ prompt-writing-assistant.json (1.4KB - Metadata only)
βββ rules-processor.md (3.7KB - Clean markdown content)
βββ rules-processor.json (1.4KB - Metadata only)
```
## π― Benefits Achieved
### β¨ Clean Content Files
- **No more JSON escaping**: Edit prompts in pure markdown
- **Readable formatting**: Proper syntax highlighting and structure
- **Version control friendly**: Clean diffs without escape characters
### π§ Separated Concerns
- **Content**: Stored in `.md` files for easy editing
- **Metadata**: Tags, variables, descriptions in `.json` files
- **Runtime combination**: Server automatically merges both
### π Example Structure (As Requested)
```typescript
// This is exactly what you wanted!
const originalPromptText = `**IMPORTANT**: This prompt is an addendum to your existing system prompt...
You are an expert software engineer and systematic problem solver...
**STEVIE'S CURRENT GOAL/REQUEST:**
{{user_goal_or_request}}`;
const promptObject = {
id: "agentic-coder-enhanced",
description: "An expert software engineer...",
content: originalPromptText, // JSON.stringify will handle this
tags: [
"agentic-coding",
"software-engineering",
"problem-solving"
],
variables: {
"user_goal_or_request": {
description: "The primary goal...",
required: true
}
},
metadata: { /* ... */ }
};
const jsonStringOutput = JSON.stringify(promptObject, null, 2);
```
## π Server Updates
### Modified `StoredPrompt` Interface
```typescript
interface StoredPrompt {
id: string;
description?: string;
content?: string; // Optional - loaded from contentFile
contentFile?: string; // Path to markdown file
tags: string[];
variables: Record<string, StoredPromptVariable>;
metadata: Record<string, unknown>;
}
```
### Enhanced File Reading
- Automatically reads markdown content when `contentFile` is specified
- Falls back to inline `content` for backward compatibility
- Ensures content is always available for template processing
## π Converted Prompts
1. **agentic-coder-enhanced** β
- 15KB markdown file with structured coding assistant content
- Variables: `user_goal_or_request`, `project_memory_file_name`
2. **memory-bank-assistant** β
- 4.7KB markdown with memory management workflows
- Variables: `rules_file`, `user_task_description`
3. **nextjs-codebase-review** β
- 3.1KB markdown with comprehensive Next.js review framework
- Variables: `project_context`
4. **python-codebase-review** β
- 3.3KB markdown with Python architecture review guidelines
- Variables: `project_context`
5. **prompt-writing-assistant** β
- 2.3KB markdown for AI prompt engineering methodology
- Variables: Multiple optional formatting parameters
6. **rules-processor** β
- 3.7KB markdown for intelligent rule processing system
- Variables: `rules_registry_path`, `file_path_or_description`, etc.
## π Mission Accomplished!
Your prompt management system is now:
- **Maintainable**: Edit markdown files directly
- **Scalable**: Add new prompts with `.md` + `.json` pairs
- **Version-control friendly**: Clean, readable diffs
- **Backward compatible**: Existing functionality unchanged
- **Developer friendly**: No more JSON escaping headaches
The server automatically combines markdown content with JSON metadata at runtime, giving you the clean structure you requested while maintaining full MCP compatibility!