reachableceo/LLMScaffolding
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

merging stuff from my joplin notes

Browse Source
This commit is contained in:
Charles N Wyble - admin
2025-07-14 09:02:35 -05:00
parent 56c223878f
commit e8b7648990
12 changed files with 1026 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
MCP/LLM Scaffolding/MCP/Basic Memory.md Executable file
View File

@@ -0,0 +1,8 @@
Basic Memory
- URLs:
- https://memory.basicmachines.co/integrations/claude-desktop
- https://github.com/basicmachines-co/basic-memory
- https://www.reddit.com/r/ClaudeAI/comments/1jdga7v/basic_memory_a_tool_that_gives_claude_persistent/
- Status: Installed/integrated with Claude Desktop

4
MCP/LLM Scaffolding/MCP/Cloudron.md Executable file
View File

@@ -0,0 +1,4 @@
Cloudron
URL: https://git.cloudron.io/playground/toy-mcp-server/
Status: cloned, deps installed, not yet hooked to Claude Desktop

26
MCP/LLM Scaffolding/MCP/Joplin.md Executable file
View File

@@ -0,0 +1,26 @@
- URLs:
- https://mcpmarket.com/server/joplin
- https://github.com/dweigend/joplin-mcp-server
- Here is the config:
```json
{
"mcpServers": {
"joplin": {
"command":"C:\\Users\\tsys\\.local\\bin\\uv.exe",
"args": [
"--directory",
"E:/tsys/LLM/MCP/joplin-mcp-server",
"run",
"src/mcp/joplin_mcp.py"
]
}
}
}
```
- Status: Configured and deployed in Claude Desktop
Here is the successful execution:
![6e5ad57ec0bb70995a2ab92d65844308.png](../../_resources/6e5ad57ec0bb70995a2ab92d65844308.png)

20
MCP/LLM Scaffolding/MCP/Redmine.md Executable file
View File

@@ -0,0 +1,20 @@
Redmine
- URL: https://github.com/runekaagaard/mcp-redmine
- Status: Configured and deployed in Claude Desktop
- Here is the config:
```json
"redmine": {
"command": "uvx",
"args": ["--from", "mcp-redmine==2025.04.09.153531",
"--refresh-package", "mcp-redmine", "mcp-redmine"],
"env": {
"REDMINE_URL": "https://projects.knownelement.com",
"REDMINE_API_KEY": "[redacted]",
"REDMINE_REQUEST_INSTRUCTIONS": "E:/tsys/LLM/MCP/redmine-mcp-server/instructions.md"
}
}
```

BIN
MCP/_resources/6e5ad57ec0bb70995a2ab92d65844308.png Executable file
View File

Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 58 KiB

164
ReleasedPrompts/LLM Scaffolding/COO Work/Doc prompt eng/Artifact Structure Guide.md Executable file
View File

@@ -0,0 +1,164 @@
# Artifact Structure Guide
Version: 1.0 (2025-03-25)
## PURPOSE
This guide establishes standard formats for creating both human-optimized and Claude-optimized artifacts to ensure consistency and effectiveness in document engineering projects.
## COMMON ELEMENTS FOR ALL ARTIFACTS
### 1. METADATA HEADER
All artifacts must include:
```
# [Artifact Title]
Version: [Version Number] ([Date in YYYY-MM-DD format])
```
### 2. VERSION CONTROL
Version numbering follows semantic versioning:
- Major version (1.0): Significant structural changes
- Minor version (0.1): Content updates that maintain structure
- Patch version (0.0.1): Small corrections or clarifications
### 3. CROSS-REFERENCES
End each artifact with related documents:
```
---
*Cross-reference with: [Related Artifact 1], [Related Artifact 2]*
```
## HUMAN-OPTIMIZED ARTIFACTS
These artifacts are designed for human readability and use.
### Structure Guidelines
1. **Hierarchical Organization**
- Use consistent header levels (H1 for title, H2 for major sections, etc.)
- Limited to 3 levels of nesting for readability
- Each level should have 2-7 subsections maximum
2. **Visual Elements**
- Use tables for comparing multiple items
- Use bulleted lists for unordered collections
- Use numbered lists for sequential instructions
- Use code blocks for templates, examples, or fill-in sections
3. **Navigation Aids**
- Include a brief purpose statement at the top
- Add a table of contents for documents over 2 pages
- Use horizontal rules to separate major sections
4. **Actionable Components**
- Include checkboxes for completion tracking
- Clearly mark required vs. optional elements
- Provide example text in italics or gray formatting
### Example Format
```markdown
# Artifact Title
Version: 1.0 (2025-03-25)
## PURPOSE
Brief description of the artifact's intended use and value.
## SECTION ONE
Main content organized with:
- Bullet points for collections
- Tables for comparisons
- Code blocks for templates
### Subsection
More detailed content with specific guidelines.
## SECTION TWO
Additional structured content.
---
*Cross-reference with: Related Artifact 1, Related Artifact 2*
```
## CLAUDE-OPTIMIZED ARTIFACTS
These artifacts are designed to be optimally processed by Claude.
### Structure Guidelines
1. **Clear Demarcation**
- Use XML-style tags to clearly indicate sections
- Include role specifications when applicable
- Separate instructions from examples with clear markers
2. **Processing Instructions**
- Place critical instructions at the beginning
- Use numbered lists for sequential processing steps
- Mark optional paths with explicit conditionals
- Include stop conditions for iterative processes
3. **Context Management**
- Group related information within single sections
- Label sections with descriptive IDs for reference
- Include summary sections for complex content
- Maintain consistent terminology throughout
4. **Output Formatting**
- Specify desired output structure explicitly
- Include sample outputs with annotations
- Define error handling and fallback options
- Provide quantitative requirements (length, detail level)
### Example Format
```
<ARTIFACT id="example_artifact" version="1.0" date="2025-03-25">
<PURPOSE>
Precisely defined purpose statement with success criteria.
</PURPOSE>
<INSTRUCTIONS>
1. First processing step with specific guidance
2. Second processing step with decision points
3. Third processing step with output requirements
</INSTRUCTIONS>
<CONTEXT id="domain_knowledge">
Essential information Claude needs to understand the domain.
</CONTEXT>
<EXAMPLES>
<EXAMPLE id="example1">
Input: Sample input
Expected output: Sample output
Explanation: Why this output meets requirements
</EXAMPLE>
</EXAMPLES>
<OUTPUT_REQUIREMENTS>
- Format specification
- Content requirements
- Quality indicators
- Error handling procedures
</OUTPUT_REQUIREMENTS>
</ARTIFACT>
```
## HYBRID ARTIFACT APPROACH
For optimal results in document engineering projects:
1. **Create human-optimized artifacts first** for team review and approval
2. **Derive Claude-optimized versions** for implementation
3. **Maintain both versions** with synchronized updates
4. **Test Claude-optimized artifacts** with various inputs
5. **Document transformation patterns** between human and Claude versions
## RECOMMENDED ARTIFACT TYPES
1. **Templates**: Structured outlines for creating consistent documents
2. **Checklists**: Sequential verification steps to ensure quality
3. **Protocols**: Standardized procedures for handling specific scenarios
4. **Repositories**: Collections of reusable components or examples
5. **Indices**: Catalogs of available artifacts with access information
---
*Cross-reference with: Terminology Bank, Validation Checklist, Comprehensive Templates Index*

183
ReleasedPrompts/LLM Scaffolding/COO Work/Doc prompt eng/Complete Workflow.md Executable file
View File

@@ -0,0 +1,183 @@
# Complete Workflow
Version: 1.0 (2025-03-25)
## PURPOSE
This workflow provides a comprehensive overview of the two-step document prompt engineering process, from initial requirement gathering through implementation and maintenance.
## TWO-STEP PROCESS OVERVIEW
The document prompt engineering process consists of two primary phases:
1. **Initial Prompt Engineering**: Focused on requirements gathering, analysis, and initial prompt creation
2. **Prompt Review**: Focused on evaluation, refinement, and optimization
This separation ensures thorough exploration followed by targeted improvement.
## DETAILED WORKFLOW
### PHASE 1: PREPARATION
#### 1.1 Project Initialization
- Define document type and purpose
- Gather sample documents
- Identify stakeholders and subject matter experts
- Establish success criteria
- Create project continuity document
#### 1.2 Requirements Analysis
- Analyze document structure and components
- Identify regulatory or compliance requirements
- Document user needs and pain points
- Map content relationships and dependencies
- Define terminology standards
#### 1.3 Environment Setup
- Create project folder structure
- Initialize artifact templates
- Establish version control approach
- Set up testing methodology
- Prepare continuity management system
### PHASE 2: INITIAL PROMPT ENGINEERING
#### 2.1 Exploratory Interview
- Conduct structured interview with Claude
- Focus on one question at a time
- Document all insights and recommendations
- Utilize extended thinking mode for complex topics
- Update continuity document with progress
#### 2.2 Document Analysis
- Analyze structural patterns
- Identify critical components
- Document domain-specific terminology
- Map content relationships
- Define document lifecycle
#### 2.3 Prompt Architecture Development
- Design context provision strategy
- Develop instruction components
- Create format specification
- Draft example blocks
- Establish guardrails and constraints
#### 2.4 Initial Prompt Creation
- Draft complete prompt using structured template
- Incorporate all key elements identified
- Ensure proper formatting and organization
- Document assumptions and design decisions
- Create Claude-optimized version if needed
#### 2.5 Preliminary Testing
- Test with sample scenarios
- Document initial performance
- Identify obvious improvement areas
- Prepare for review session
- Update continuity document
### PHASE 3: PROMPT REVIEW
#### 3.1 Performance Assessment
- Evaluate against success criteria
- Analyze output quality and consistency
- Identify strengths and weaknesses
- Document performance metrics
- Prioritize improvement areas
#### 3.2 Structural Analysis
- Evaluate context effectiveness
- Assess instruction clarity
- Review format adherence
- Analyze example utility
- Document structural insights
#### 3.3 Content Quality Evaluation
- Assess factual accuracy
- Evaluate completeness
- Review logical flow
- Check tone appropriateness
- Verify terminology consistency
#### 3.4 Optimization Strategy
- Develop targeted improvements
- Create experimental variations
- Design comparison methodology
- Define success indicators
- Document optimization approach
#### 3.5 Prompt Revision
- Implement prioritized improvements
- Create revised prompt draft
- Document changes and rationale
- Update related artifacts
- Prepare for validation testing
#### 3.6 Validation Testing
- Test with predefined scenarios
- Compare to original performance
- Document improvements
- Identify any new issues
- Make final adjustments
### PHASE 4: IMPLEMENTATION
#### 4.1 Finalization
- Make go/no-go decision
- Finalize prompt documentation
- Create implementation guide
- Establish monitoring plan
- Update all project artifacts
#### 4.2 Deployment
- Implement in target environment
- Provide user training if needed
- Monitor initial performance
- Gather feedback
- Document deployment process
#### 4.3 Performance Monitoring
- Track key performance indicators
- Collect user feedback
- Document edge cases
- Identify optimization opportunities
- Schedule periodic reviews
### PHASE 5: MAINTENANCE
#### 5.1 Periodic Review
- Evaluate ongoing performance
- Identify changing requirements
- Update prompt as needed
- Document review findings
- Schedule next review
#### 5.2 Continuous Improvement
- Implement minor optimizations
- Test variations
- Document performance changes
- Update related artifacts
- Maintain version history
## ROLE-SPECIFIC WORKFLOWS
### FOR PROMPT ENGINEERS
1. Lead initial requirements gathering
2. Conduct exploratory interviews with Claude
3. Draft initial prompt architecture
4. Create example blocks
5. Document design decisions
6. Participate in review sessions
7. Implement technical optimizations
8. Update technical documentation
### FOR SUBJECT MATTER EXPERTS
1. Provide domain knowledge
2. Review document components for accuracy
3. Validate terminology usage
4. Assess output quality
5. Identify domain-specific edge cases
6. Help prioritize improvements
7. Review final documentation
8. Support

116
ReleasedPrompts/LLM Scaffolding/COO Work/Doc prompt eng/Continuity Document.md Executable file
View File

@@ -0,0 +1,116 @@
# Continuity Document
Version: 1.0 (2025-03-25)
## PURPOSE
This master document serves as a persistent record of project progress and current status, enabling seamless resumption of work across multiple sessions and different Claude instances.
## PROJECT METADATA
```
[PROJECT TITLE]:
[START DATE]:
[CURRENT PHASE]:
[TARGET COMPLETION DATE]:
[KEY STAKEHOLDERS]:
[PRIMARY OBJECTIVES]:
```
## CURRENT STATUS SUMMARY
Provide a concise overview of the current state of the project (50-75 words).
```
[CURRENT STATUS]:
```
## ACTIVE ARTIFACTS INVENTORY
List all artifacts currently in use with their latest versions.
| Artifact Name | Version | Last Updated | Description |
|---------------|---------|--------------|-------------|
| [Artifact 1] | [v#.#] | [YYYY-MM-DD] | [Brief description] |
| [Artifact 2] | [v#.#] | [YYYY-MM-DD] | [Brief description] |
| [Add more rows as needed] |
## CONVERSATION TIMELINE
### Session 1: [Date]
- Key accomplishments:
- [Accomplishment 1]
- [Accomplishment 2]
- Decisions made:
- [Decision 1]
- [Decision 2]
- Created artifacts:
- [Artifact 1] (v1.0)
- [Artifact 2] (v1.0)
### Session 2: [Date]
- Key accomplishments:
- [Accomplishment 1]
- [Accomplishment 2]
- Decisions made:
- [Decision 1]
- [Decision 2]
- Updated artifacts:
- [Artifact 1] (v1.1): [Summary of changes]
- [Artifact 3] (v1.0): [New artifact]
### [Add more sessions as needed]
## CURRENT FOCUS
```
[CURRENT TASK]:
[PROGRESS ON CURRENT TASK]:
[BLOCKING ISSUES]:
[NEXT IMMEDIATE ACTIONS]:
```
## DECISIONS LOG
Record all significant decisions with rationale.
| Date | Decision | Rationale | Alternatives Considered |
|------|----------|-----------|-------------------------|
| [YYYY-MM-DD] | [Decision summary] | [Brief rationale] | [Alternative approaches] |
| [Add more rows as needed] |
## ASSUMPTIONS REGISTER
Document all assumptions being made in the project.
| ID | Assumption | Impact if Invalid | Verification Method |
|----|------------|-------------------|---------------------|
| A1 | [Assumption description] | [Potential impact] | [How/when to verify] |
| [Add more rows as needed] |
## CONVERSATION CONTEXT PRESERVATION
Key information to maintain conversation continuity:
```
[RECENT DISCUSSION POINTS]:
[PENDING QUERIES]:
[UNRESOLVED QUESTIONS]:
[AGREED TERMINOLOGY]:
```
## NEXT STEPS
- [ ] [Next action 1]
- [ ] [Next action 2]
- [ ] [Next action 3]
- [ ] [Additional actions as needed]
## TRANSITION NOTES
Information specifically for Claude when resuming in a new session:
```
[CRITICAL CONTEXT]:
[SPECIALIZED ROLES/EXPERTISE NEEDED]:
[PREFERRED COMMUNICATION STYLE]:
[ADDITIONAL INSTRUCTIONS]:
```
## APPENDIX: SESSION RESUMPTION INSTRUCTIONS
To resume work on this project in a new session, please:
1. Share this continuity document with Claude
2. Use the prompt: "Document engineering chat - Please review this continuity document and help me resume our work exactly where we left off."
3. Confirm that Claude acknowledges the current status and next steps
4. Proceed with the next action items as outlined above
---
*Cross-reference with: Multi-Session Handoff, Chat Transition Protocol, Comprehensive Templates Index*

167
ReleasedPrompts/LLM Scaffolding/COO Work/Doc prompt eng/Exploratory Interview Guide.md Executable file
View File

@@ -0,0 +1,167 @@
# Exploratory Interview Guide
Version: 1.0 (2025-03-25)
## PURPOSE
This guide provides a structured approach to conducting exploratory interviews with Claude during document prompt engineering, emphasizing a one-question-at-a-time methodology to maximize the effectiveness of extended thinking mode.
## CORE PRINCIPLES
1. **Hierarchical Questioning**: Move from general to specific
2. **Active Listening**: Adapt questions based on previous responses
3. **Single-Focus Queries**: One concept per question
4. **Deliberate Progression**: Build a logical path of inquiry
5. **Extended Thinking Activation**: Strategic use of complex questions
## QUESTION HIERARCHY FRAMEWORK
### Level 1: Domain Understanding
These questions establish fundamental context and document characteristics.
**Example Questions:**
- "What are the defining characteristics of [document type]?"
- "What purpose does [document type] serve in [specific industry]?"
- "What components are essential for a complete [document type]?"
- "What common errors or omissions occur in [document type]?"
- "How has [document type] evolved over the past 5 years?"
**Follow-up Pattern:**
- Ask for clarification on unexpected elements
- Request examples of mentioned characteristics
- Explore regional or industry variations
### Level 2: Structure Analysis
These questions examine the format and organization of the document.
**Example Questions:**
- "What organizational structure is most effective for [document type]?"
- "How should information hierarchy be established in [document type]?"
- "What sections are mandatory vs. optional in [document type]?"
- "How does the structure of [document type] support its usability?"
- "What dependencies exist between different sections of [document type]?"
**Follow-up Pattern:**
- Probe for reasoning behind structural recommendations
- Ask about alternative structures and their tradeoffs
- Explore how structure affects document effectiveness
### Level 3: Content Optimization
These questions focus on the quality and effectiveness of the document's content.
**Example Questions:**
- "What language patterns increase clarity in [document type]?"
- "How should technical terminology be handled in [document type]?"
- "What level of detail is appropriate for [specific section]?"
- "How can [document type] balance comprehensiveness with readability?"
- "What contextual information is necessary for [specific element]?"
**Follow-up Pattern:**
- Request examples of optimal vs. suboptimal content
- Explore content variations for different audiences
- Ask about content evolution over document lifecycle
### Level 4: Edge Case Management
These questions address unusual scenarios and exception handling.
**Example Questions:**
- "How should [document type] address [uncommon scenario]?"
- "What contingencies should be included for [potential issue]?"
- "What flexibility is required in [document type] to accommodate [variable factor]?"
- "How can [document type] maintain integrity when [challenging condition]?"
- "What failsafes should be incorporated into [document type]?"
**Follow-up Pattern:**
- Explore frequency and impact of edge cases
- Ask about prioritization of different edge cases
- Request examples of well-handled edge cases
### Level 5: Implementation Strategy
These questions focus on practical application and execution.
**Example Questions:**
- "What approach would you recommend for transitioning from [current document] to [optimized document]?"
- "How should [specific challenge] be addressed during implementation?"
- "What metrics would effectively measure the success of [document type]?"
- "What timeline is realistic for developing and implementing [document type]?"
- "How should feedback be incorporated into [document type] iterations?"
**Follow-up Pattern:**
- Ask for step-by-step implementation recommendations
- Explore potential obstacles and solutions
- Request examples of successful implementations
## EXTENDED THINKING MODE ACTIVATION
For optimal use of Claude's extended thinking mode, structure questions that:
1. **Require Multivariate Analysis**
- "How would changes to [element X] impact [components A, B, and C] in [document type]?"
- "What interdependencies exist between [element X] and [elements Y and Z]?"
2. **Involve Complex Tradeoffs**
- "What are the tradeoffs between [approach A] and [approach B] when considering [factors X, Y, and Z]?"
- "How should [competing priorities X and Y] be balanced in [document type]?"
3. **Request Synthesized Insights**
- "Based on our discussion of [elements A, B, and C], what overarching principles should guide [document type]?"
- "What patterns emerge when comparing [scenarios X, Y, and Z] in [document type]?"
4. **Require Judgment in Ambiguous Scenarios**
- "Given the uncertainty around [factor X], what approach would you recommend for [document element]?"
- "How should [document type] handle situations where [conflicting requirements] exist?"
## FOLLOW-UP PROTOCOLS
### 1. Clarification Sequence
When Claude's response requires clarification:
1. Echo the unclear point: "You mentioned [specific point]. Could you clarify what you mean by that?"
2. Ask for an example: "Could you provide an example of [unclear concept]?"
3. Request reframing: "Could you explain [unclear concept] in a different way?"
4. Offer interpretation: "I understand [unclear concept] to mean [your interpretation]. Is that correct?"
### 2. Expansion Sequence
When Claude's response should be expanded:
1. Request depth: "Could you elaborate further on [specific aspect]?"
2. Ask for additional dimensions: "Beyond [mentioned aspects], what other factors should be considered?"
3. Request implications: "What are the implications of [specific aspect] for [related element]?"
4. Explore variations: "How might [specific aspect] vary in different contexts?"
### 3. Contrast Sequence
When comparing approaches or elements:
1. Request direct comparison: "How does [approach A] compare to [approach B] regarding [specific criterion]?"
2. Ask for advantages/disadvantages: "What are the pros and cons of [approach] compared to alternatives?"
3. Explore contextual variations: "In what scenarios would [approach A] be preferable to [approach B]?"
4. Request integration possibilities: "Could elements of [approach A] and [approach B] be combined?"
### 4. Validation Sequence
When confirming understanding or alignment:
1. Summarize understanding: "Based on our discussion, my understanding is [summary]. Is that accurate?"
2. Test with scenarios: "If [specific scenario] occurred, how would the approach you've described address it?"
3. Challenge assumptions: "What assumptions underlie the approach you've described?"
4. Explore limitations: "What limitations or constraints should we be aware of with this approach?"
## DOCUMENTATION PRACTICES
For each question and answer:
1. Record the exact question asked
2. Document Claude's complete response
3. Note any areas requiring follow-up
4. Track key insights and recommendations
5. Update the project continuity document with session progress
## INTERVIEW TERMINATION CRITERIA
Conclude the interview when:
- All planned question topics have been covered
- Response patterns become repetitive without new insights
- Clear recommendations have emerged for all key areas
- Sufficient material has been gathered for prompt development
- Chat context length approaches limitations
---
*Cross-reference with: Initial Prompt Engineering Session, Validation Checklist, Chat Transition Protocol*

109
ReleasedPrompts/LLM Scaffolding/COO Work/Doc prompt eng/Initial Prompt Engineering Session Template.md Executable file
View File

@@ -0,0 +1,109 @@
# Initial Prompt Engineering Session Template
Version: 1.0 (2025-03-25)
## PURPOSE
This template guides the initial session of document prompt engineering to systematically develop an effective prompt structure for any document type.
## PREPARATION CHECKLIST
- [ ] Document samples collected (minimum 3 recommended)
- [ ] Target outcome defined
- [ ] Stakeholder requirements documented
- [ ] Domain expertise identified/secured
- [ ] Technical constraints acknowledged
## SESSION STRUCTURE
### 1. DOCUMENT ANALYSIS
```
[DOCUMENT TYPE]:
[PRIMARY PURPOSE]:
[TARGET AUDIENCE]:
[CRITICAL COMPONENTS]:
[DOMAIN-SPECIFIC TERMINOLOGY]:
[STRUCTURAL PATTERNS]:
[PAIN POINTS IN CURRENT PROCESS]:
```
### 2. PROMPT OBJECTIVES
```
[PRIMARY OBJECTIVE]:
[SECONDARY OBJECTIVES]:
[SUCCESS CRITERIA]:
[FAILURE MODES TO AVOID]:
```
### 3. PROMPT ARCHITECTURE
#### 3.1 Context Provision
```
[ESSENTIAL BACKGROUND]:
[DOMAIN KNOWLEDGE REQUIREMENTS]:
[REGULATORY/COMPLIANCE ELEMENTS]:
```
#### 3.2 Instruction Components
```
[CORE INSTRUCTIONS]:
[SUPPLEMENTARY GUIDELINES]:
[GUARDRAILS AND CONSTRAINTS]:
```
#### 3.3 Format Specification
```
[REQUIRED SECTIONS]:
[STRUCTURAL REQUIREMENTS]:
[STYLISTIC GUIDELINES]:
```
#### 3.4 Examples Block
```
[INPUT EXAMPLE 1]:
[EXPECTED OUTPUT 1]:
[INPUT EXAMPLE 2]:
[EXPECTED OUTPUT 2]:
```
### 4. QUESTION SEQUENCE FOR CLAUDE
Use the Exploratory Interview Guide artifact for structured questioning hierarchy. Document each question and answer from Claude.
```
Q1: [First high-level question]
A1: [Claude's response]
Q2: [Follow-up based on response]
A2: [Claude's response]
[CONTINUE SEQUENCE]
```
### 5. INITIAL PROMPT DRAFT
```
[FULL PROMPT TEXT]:
```
### 6. PRELIMINARY TESTING NOTES
```
[TEST SCENARIO 1]:
[OUTCOME]:
[OBSERVATIONS]:
[TEST SCENARIO 2]:
[OUTCOME]:
[OBSERVATIONS]:
```
### 7. NEXT STEPS
- [ ] Schedule prompt review session
- [ ] Identify subject matter experts for review
- [ ] Plan A/B testing methodology
- [ ] Update continuity document with session outcomes
## NOTES
- Use extended thinking mode for complex document analysis
- Maintain consistent terminology (reference Terminology Bank artifact)
- Document all assumptions made during the session
- Flag any areas requiring additional research
---
*Cross-reference with: Artifact Structure Guide, Exploratory Interview Guide, Terminology Bank*

122
ReleasedPrompts/LLM Scaffolding/COO Work/Doc prompt eng/Prompt Review Session Template.md Executable file
View File

@@ -0,0 +1,122 @@
# Prompt Review Session Template
Version: 1.0 (2025-03-25)
## PURPOSE
This template facilitates a structured review of prompts developed during the initial engineering session to refine and optimize their effectiveness.
## PREPARATION CHECKLIST
- [ ] Initial prompt draft ready for review
- [ ] Test results from initial implementation documented
- [ ] Review team assembled with appropriate expertise
- [ ] Success criteria clearly defined
- [ ] Previous session notes accessible
## SESSION STRUCTURE
### 1. PROMPT PERFORMANCE ASSESSMENT
```
[PROMPT UNDER REVIEW]:
[INITIAL SUCCESS CRITERIA]:
[PERFORMANCE METRICS]:
[AREAS MEETING EXPECTATIONS]:
[AREAS REQUIRING IMPROVEMENT]:
```
### 2. STRUCTURAL ANALYSIS
#### 2.1 Context Effectiveness
```
[SUFFICIENT BACKGROUND PROVIDED?]: Yes/No
[MISSING CONTEXT ELEMENTS]:
[SUPERFLUOUS CONTEXT ELEMENTS]:
[CLARITY OF DOMAIN KNOWLEDGE]:
```
#### 2.2 Instruction Clarity
```
[AMBIGUOUS INSTRUCTIONS IDENTIFIED]:
[CONFLICTING GUIDELINES IDENTIFIED]:
[INSTRUCTION HIERARCHY CLEAR?]: Yes/No
[GUARDRAIL EFFECTIVENESS]:
```
#### 2.3 Format Adherence
```
[FORMAT REQUIREMENTS MET?]: Yes/No
[STRUCTURAL INCONSISTENCIES]:
[STYLISTIC DEVIATIONS]:
[QUALITY OF OUTPUT STRUCTURE]:
```
#### 2.4 Example Utility
```
[EXAMPLES PROPERLY REFERENCED?]: Yes/No
[EXAMPLE DIVERSITY SUFFICIENT?]: Yes/No
[MISSING EXAMPLE SCENARIOS]:
```
### 3. CONTENT QUALITY EVALUATION
```
[FACTUAL ACCURACY]:
[COMPLETENESS OF CONTENT]:
[LOGICAL FLOW]:
[APPROPRIATE TONE]:
[TERMINOLOGY CONSISTENCY]:
```
### 4. OPTIMIZATION STRATEGIES
#### 4.1 Targeted Improvements
```
[SPECIFIC ELEMENT]:
[CURRENT STATE]:
[PROPOSED REVISION]:
[RATIONALE]:
```
#### 4.2 Experimental Variations
```
[VARIATION A]:
[VARIATION B]:
[COMPARISON METHODOLOGY]:
[SUCCESS INDICATORS]:
```
### 5. REVISED PROMPT
```
[UPDATED FULL PROMPT TEXT]:
```
### 6. VALIDATION TESTING
```
[TEST SCENARIO 1]:
[OUTCOME]:
[IMPROVEMENT FROM ORIGINAL?]: Yes/No/Partial
[TEST SCENARIO 2]:
[OUTCOME]:
[IMPROVEMENT FROM ORIGINAL?]: Yes/No/Partial
```
### 7. IMPLEMENTATION RECOMMENDATIONS
```
[GO/NO-GO DECISION]:
[RECOMMENDED DEPLOYMENT CONTEXT]:
[MONITORING REQUIREMENTS]:
[EXPECTED MAINTENANCE FREQUENCY]:
```
### 8. NEXT STEPS
- [ ] Finalize prompt documentation
- [ ] Schedule follow-up review if needed
- [ ] Plan implementation strategy
- [ ] Update continuity document with review outcomes
## NOTES
- Use validation checklist to prevent redundant testing
- Document all stakeholder feedback received
- Consider A/B testing for significant changes
- Flag edge cases for special handling
---
*Cross-reference with: Validation Checklist, Initial Prompt Engineering Session, Continuity Document*

107
ReleasedPrompts/LLM Scaffolding/COO Work/Doc prompt eng/Terminology Bank.md Executable file
View File

@@ -0,0 +1,107 @@
# Terminology Bank
Version: 1.0 (2025-03-25)
## PURPOSE
This terminology bank establishes standardized definitions for terms used in document prompt engineering to ensure consistency across artifacts and conversations.
## USAGE INSTRUCTIONS
1. Refer to this bank when introducing new terms in artifacts
2. Update with new terminology as it emerges in the project
3. Resolve any conflicting definitions before proceeding
4. Include relevant domain-specific terms for each project
5. Reference in the continuity document to maintain consistency across sessions
## CORE TERMINOLOGY
### Document Engineering Concepts
| Term | Definition | Usage Notes |
|------|------------|-------------|
| Document Prompt Engineering | The systematic process of developing, testing, and refining prompts specifically for document creation or analysis. | Preferred over "prompt design" or "prompt crafting" |
| Initial Engineering Session | The first phase of prompt development focused on understanding requirements and drafting initial approaches. | Always followed by at least one review session |
| Prompt Review Session | A structured evaluation of prompt performance with specific optimization objectives. | May be iterative depending on complexity |
| Artifact | A self-contained document that serves a specific function in the prompt engineering process. | All artifacts should follow the structure guide |
| Two-Step Process | The complete document prompt engineering workflow consisting of initial engineering and subsequent review. | Standard approach for all document types |
| Continuity Management | Techniques to maintain project context across multiple sessions or Claude instances. | Essential for complex projects |
### Prompt Structure Elements
| Term | Definition | Usage Notes |
|------|------------|-------------|
| Context Block | Information provided to Claude about the document domain, purpose, and background. | Should be comprehensive but concise |
| Instruction Set | Specific directives for Claude regarding document creation or analysis. | Organized hierarchically by priority |
| Format Specification | Requirements for the structure, style, and organization of the output. | Includes both mandatory and optional elements |
| Examples Block | Sample inputs and outputs that illustrate desired performance. | Should cover standard cases and edge cases |
| Guardrails | Constraints and limitations that prevent undesired outputs. | Both positive and negative constraints |
| Extended Thinking Trigger | Phrasings designed to activate Claude's extended reasoning capabilities. | Used for complex analytical requirements |
### Document Components
| Term | Definition | Usage Notes |
|------|------------|-------------|
| Structural Element | Any component that defines the organization of a document. | E.g., sections, subsections, headers |
| Content Element | Text, data, or media that provides information within the document. | Distinguished from structural elements |
| Conditional Element | Document components that appear only when specific criteria are met. | Important for adaptable documents |
| Metadata | Information about the document itself rather than its subject matter. | E.g., version, author, date, status |
| Compliance Element | Components required to meet regulatory or policy requirements. | Should be clearly identified as mandatory |
| User Input Field | Designated area for information to be provided by end-users. | Requires clear instructions and validation |
### Process Terminology
| Term | Definition | Usage Notes |
|------|------------|-------------|
| Exploratory Interview | A structured conversation with Claude to gather insights for prompt development. | Follows the interview guide artifact |
| Validation Testing | Systematic evaluation of prompt performance against predefined criteria. | Uses the validation checklist |
| A/B Testing | Comparison of alternative prompt versions to determine optimal approach. | Requires consistent test scenarios |
| Prompt Iteration | The process of refining a prompt based on performance analysis. | Should be tracked in the continuity document |
| Chat Transition | The process of moving a conversation to a new chat when approaching length limitations. | Follows the transition protocol |
| Multi-Session Handoff | The process of resuming work across different conversations. | Uses the handoff template |
## PROJECT-SPECIFIC TERMINOLOGY
Add domain-specific terminology relevant to the current project.
### [Domain Name]
| Term | Definition | Usage Notes |
|------|------------|-------------|
| [Term 1] | [Definition] | [Usage notes] |
| [Term 2] | [Definition] | [Usage notes] |
| [Add more as needed] |
### [Additional Domains as Needed]
| Term | Definition | Usage Notes |
|------|------------|-------------|
| [Term 1] | [Definition] | [Usage notes] |
| [Term 2] | [Definition] | [Usage notes] |
| [Add more as needed] |
## TERMINOLOGY MAINTENANCE
### Addition Process
To add new terminology:
1. Confirm term is not already defined (with potential variations)
2. Draft clear, concise definition
3. Provide usage notes with examples
4. Update the version number of this document
5. Reference the update in the continuity document
### Revision Process
To revise existing terminology:
1. Document both old and new definitions
2. Provide rationale for the change
3. Update all artifacts using the term
4. Note the revision in the continuity document
5. Update the version number of this document
### Retirement Process
To retire obsolete terminology:
1. Mark as deprecated with date
2. Provide recommended alternative term(s)
3. Maintain in the bank with deprecated status for reference
4. Update all artifacts using the term
5. Update the version number of this document
---
*Cross-reference with: Artifact Structure Guide, Continuity Document, Comprehensive Templates Index*