Prompts
Prompts
All prompts in this system follow a standardized structure defined by the Prompt Template. This template ensures consistency across all workflow prompts with a clear and predictable format.
Standardized Template Structure
Every prompt follows this four-part structure:
Structure Components
Input
- Required inputs, file paths, data sources, and parameters
- Clearly defines what the prompt needs to work with
- Specifies any optional parameters or configurations
Output
- Expected output format and structure
- File naming conventions
- Location where outputs should be saved
- Any side effects or state changes
Main Process
- Step-by-step workflow in numbered format
- Sub-bullets for detailed steps within each main step
- Clear, actionable instructions for execution
Caveats
- Critical warnings and important considerations
- Processing rules and constraints
- Quality standards and best practices
- Edge cases to be aware of
Benefits
The standardized prompt template provides several key advantages:
- Consistent Documentation: All prompts follow the same structure, making them easier to understand and use
- Clear Specifications: Input/output requirements are explicitly defined for each workflow step
- Easier Maintenance: Updates and improvements can be made systematically
- Better Understanding: Users and AI agents can quickly grasp prompt requirements and constraints
- Reduced Errors: Explicit caveats and processing rules prevent common mistakes
Prompt Directory Structure
All prompts are organized in the vault:
_Settings_/
├── Prompts/
│ ├── Generate Daily Roundup (GDR).md
│ ├── Generate Weekly Roundup (GWR).md
│ ├── Process Life Logs (PLL).md
│ ├── Enrich Ingested Content (EIC).md
│ ├── Topic Knowledge Creation (TKC).md
│ ├── Topic Knowledge Improvement (TKI).md
│ ├── Topic Knowledge Addendum (TKA).md
│ ├── Update Folder Notes (UFN).md
│ ├── Knowledge Task Generator (KTG).md
│ ├── Knowledge Task Processor (KTP).md
│ ├── Knowledge Task Evaluator (KTE).md
│ ├── Pick and Process Photos (PPP).md
│ ├── Create Thread Postings (CTP).md
│ ├── Ad-hoc Research within PKM (ARP).md
│ └── Adhoc/
│ └── custom_prompt.md
└── Templates/
├── Prompt Template.md
├── Journal Template.md
├── Topic Template.md
└── Weekly Roundup Template.md
Key Prompts
Daily Operations
- Generate Daily Roundup (GDR): Creates daily summaries linking meaningful updates and topics
- Process Life Logs (PLL): Summarizes voice-based life logs with memorable items and emotions
- Pick and Process Photos (PPP): Processes photos into structured photo logs
- Enrich Ingested Content (EIC): Improves captured content structure and adds summaries
Weekly Operations
- Generate Weekly Roundup (GWR): Generates weekly summaries from daily roundups
- Weekly Reading Review: Reviews weekly reading for key learnings
Topic Management
- Topic Knowledge Creation (TKC): Creates new topic notes from source materials
- Topic Knowledge Improvement (TKI): Improves topic note structure and balance
- Topic Knowledge Addendum (TKA): Updates existing topic notes with new content
Maintenance
- Update Folder Notes (UFN): Creates and updates folder notes with recent changes
- Knowledge Task Generator (KTG): Generates maintenance and improvement tasks
- Knowledge Task Processor (KTP): Processes and executes knowledge tasks
- Knowledge Task Evaluator (KTE): Evaluates task completion and quality
Content Creation
- Create Thread Postings (CTP): Generates social media content from knowledge base
- Ad-hoc Research within PKM (ARP): Conducts research within your knowledge base
Usage
Prompts can be executed in several ways:
- Via Claude Code/Cursor: Directly reference prompts in your vault
- Via CLI: Use the ai4pkm CLI tool with prompt names or inline prompts
- Manually: Follow the prompt steps in the markdown files
Example CLI usage:
# Execute a named workflow
ai4pkm -p "DIR for today"
# Execute a specific prompt inline
ai4pkm -p "GDR for yesterday"
# Use a specific agent for a prompt
ai4pkm -a gemini -p "EIC on new articles"
See Also
- Guidelines - PKM overview and principles
- Workflows - How prompts work together in workflows
- FAQ - Frequently asked questions