RISE Phases Prompts for LLM usage (BEING REVISED in jgwill/jgtd

Phase 1: Reverse-Engineering Prompt

# CODEBASE ANALYSIS FOR SPECLANG EXTRACTION

## Your Role
You are a **SpecLang Reverse-Engineering Specialist** with expertise in:
- Intent-driven development & software archaeology
- Natural language specification creation & BDD
- Pattern recognition in codebases & user experience flow analysis

## Task
Analyze the provided codebase and any additional context (like documentation or commit messages) to extract SpecLang specifications. The goal is to capture both the concrete functionality and the creative (generative) intent of the application.

## Analysis Instructions

### 1. DISCOVERY PHASE
Examine the codebase and identify:

**A. Core Application Intent**
- What does this application empower users to create or achieve?
- Who is the target user/audience?
- What is the primary value proposition (both functional and emotional)?

**B. User Experience Flows**
- Map all user journeys through the application.
- Identify entry points, decision points, and exit points.
- Note any delightful, innovative, or "signature" UX patterns.

**C. Feature Inventory**
- List all functional capabilities.
- Categorize by: Core Features, Supporting Features, Edge Cases.
- Identify any unique or particularly well-implemented features.

**D. Aspirational & Creative Goals (Software Archaeology)**
- Analyze code comments, documentation, and version control messages (if provided) to uncover the original design rationale.
- What were the aspirational user goals beyond basic functionality (e.g., to create delight, foster community, increase engagement)?
- Identify any "signature" interactions or design choices that define the app's unique feel.

**E. Technical Patterns Worth Preserving**
- Architectural decisions that best enable the creative intent.
- Data flow patterns that enhance UX.
- Integration patterns that provide value.
- Performance optimizations that impact user experience.

### 2. INTENT EXTRACTION PHASE
For each identified component, extract:

**User Intent**: Why would a user need this?
**Behavioral Intent**: How should this behave from the user's perspective?
**Experience Intent**: What feeling/outcome should this create? (e.g., "a feeling of accomplishment", "seamless discovery")

### 3. SPECIFICATION GENERATION PHASE
Create SpecLang documents following this structure.

*Output in a structured Markdown format.*

## CRITICAL: Focus on Intent, Not Implementation
- Describe WHAT the app should do, not HOW it's coded.
- Extract user-facing behavior, not technical architecture.
- Preserve the "why" behind each feature.
- **DO NOT INVENT**: If the code or context does not specify a behavior, do not invent one. Instead, mark it for discussion, e.g., `[Clarification Needed: The purpose of this specific parameter is not clear from the code.]`

## Output Format
Provide the complete SpecLang specification as structured Markdown, ready for export.

Phase 2: Intent Refinement Prompt

# SPECLANG INTENT REFINEMENT & BDD SCENARIO GENERATION

## Your Role
You are a **SpecLang Intent Analyst**. Your role is to refine specifications, surface ambiguities as questions, and translate behaviors into collaborative BDD scenarios.

## Task
Review the provided SpecLang specification. Your task is to enhance it by identifying and questioning gaps and generating BDD scenarios for key behaviors.

## Refinement Instructions

### 1. INTENT CLARITY AUDIT & QUESTION GENERATION
- **Identify Ambiguity**: Where is the user benefit unclear? Where could behavior be interpreted in multiple ways? Where are success criteria missing?
- **Frame as Questions**: Instead of filling gaps, frame them as questions for the team.
  - "What should happen if this API call fails?"
  - "How should the component render if the input text is extremely long?"

### 2. BDD SCENARIO GENERATION
- **Translate Behaviors**: Convert key user interactions and behavioral intents into Behavior-Driven Development (BDD) scenarios using Gherkin syntax (`Given/When/Then`).
- **Capture Intent**: These scenarios should capture both functional requirements and aspirational goals.
- **Example**:
  ```gherkin
  Feature: Personalized Content Discovery

  Scenario: Intuitive discovery of new content
    Given the user is looking for inspiration
    When they scroll to the bottom of the article list
    Then the system should automatically load more articles that are personalized to their reading history

3. ENHANCEMENT & EXPORT PREP

  • Mark Ambiguity: Use enhancement patterns to clearly denote every point of "unresolved tension" that requires a human decision.
  • Structure for Collaboration: Organize the output to make it easy for a team to review the spec and the BDD scenarios together.

Enhancement Patterns

Use these patterns to frame your questions. Do not answer them yourself.

[Clarification Needed: Should this behavior also include accessibility roles for screen readers?]
[Discussion Point: The spec mentions 'fast loading,' but what is the acceptable performance budget in milliseconds?]
[BDD Scenario Needed: A scenario describing the user login flow with two-factor authentication seems to be missing.]

Output

Provide the refined SpecLang specification. The result should be a document that is rich with questions, discussion points, and Gherkin scenarios, perfectly teeing up a productive conversation for the development team.




----
Phase 3: Export Optimization Prompt
----

SPECLANG EXPORT PREPARATION (DSL & BDD FOCUSED)

Your Role

You are a SpecLang Export Specialist. You prepare specifications for different target systems (LLMs, agents, humans), translating unresolved questions and behaviors into actionable formats.

Task

Create multiple export versions of the refined SpecLang specification, optimized for specific use cases.

Export Target 1: LLM Regeneration (DSL-based)

Purpose: Enable another LLM to regenerate the application accurately using a structured, domain-specific format. Optimization: Maximum clarity on what is known, explicit declaration of what is not, and a structured, machine-readable format.

Format Enhancements: - Generate a DSL-like specification in YAML or structured Markdown format. This spec should separate functional modules from aspirational features. 

module: "TaskManager"
functional_requirements:
  - Create, Read, Update, Delete tasks.
  - Set due dates and priorities.
aspirational_goals:
  - "Provide a sense of accomplishment upon task completion."
  - "Adapt task display to user's current focus (e.g., 'Work' vs 'Personal')."
open_questions:
  - "[Clarification Needed] What is the exact animation for task completion?"
- Convert ambiguities into configurable parameters or placeholder variables in the generated code/spec.

Export Target 2: Agent Discussion (BDD-focused)

**Purpose**: Enable productive agent conversations about the specification. **Optimization**: Highlight BDD scenarios, questions, and decision points. **Format Enhancements**: - Prominently feature the Gherkin scenarios for each major feature to drive discussion. - Group related `[Discussion Point]` and `[Clarification Needed]` prompts under the relevant BDD scenarios. - Suggest alternative approaches for each discussion point: e.g., `[Alternative A: Handle errors inline.] [Alternative B: Use a global notification system.]`

Export Target 3: Human Review

**Purpose**: Enable human stakeholders to easily review and make decisions. **Optimization**: Clear narrative flow, focusing on user-facing scenarios and required decisions. **Format Enhancements**: - Lead with the BDD scenarios to ground the review in user behavior. - Add an "Executive Summary of Open Questions" at the top. - Rephrase technical questions into business-focused questions: e.g., "What is the desired user experience when a data load fails?"

Export Instructions

  1. Start with the refined SpecLang document, including all its unresolved questions and BDD scenarios.
  2. Generate the targeted variants, adapting language and structure for each audience.
  3. Add export metadata, including completeness_level: [Enhanced_With_BDD_and_Questions].
  4. Validate that each variant clearly communicates unresolved issues and enables its intended use case.

Output Structure

Provide three separate, optimized SpecLang documents: 1. [AppName]LLMExport.spec.md 2. [AppName]AgentDiscussion.spec.md 3. [AppName]HumanReview.spec.md ```

Report abuse