Prompt.system.deep Dive
# Prompt System Architecture
## Status
- Version: v0 (MVP)
- Stability: stable for single-user local usage
- Concurrency: optimistic (builder_content only)
## Scope
This document explains:
- data model
- FS + DB sync
- prompt lifecycle (create, read, update)
## Gotchas
- FS must NEVER be written before DB conflict check
- Paths are stored, not inferred
- output.json must always be valid JSON
- builder_content is stored as string (JSON.stringify)
- DB is source of truth for structure, not FS
## 1. Core Mental Model
* **Prompt = container**
* **Version = snapshot of data**
* **Filesystem (FS) = actual content**
* **Database (DB) = structure + metadata**
---
## 2. Data Layer Design
### Tables
#### prompts (container)
* id
* name
* template_id
* collection_id
* current_version_id
* created_at
* updated_at
#### prompt_versions (actual data)
* id
* prompt_id
* version_number
* label
* builder_content (JSON string)
* scratchpad_path
* output_path
* output_type (optional)
* created_at
* updated_at
#### collections (folders)
* id
* name
* parent_id (nested folders)
---
## 3. Filesystem Design
### Structure
Initial (no versioning):when to use it
Community prompt sourced from the open-source GitHub repo byhsr/prmptly (no explicit license). A "Prompt.system.deep Dive" style prompt — adapt the placeholders and specifics to your task. Imported as-is and not independently retested here, so check the output before relying on it.
tags
roleplaycommunitygeneral
source
byhsr/prmptly · no explicit license