story-outline

description: Analyze story specification and identify technical decisions that need research, preparing for planning phase.

The user input to you can be provided directly by the agent or as a command argument - you MUST consider it before proceeding with the prompt (if not empty).

User input:

$ARGUMENTS

1. •

   Locate the active story directory (STORY_DIR):

   • •Preferred: user is already in the story directory (or a subfolder like contracts/)
   • •Determine STORY_DIR by walking up from current directory until you find spec.md
   • •If no spec.md found: ask the user to cd into the story directory or run /speck to route
2. •

   Load and analyze the story specification:

   • •Read {STORY_DIR}/spec.md to understand requirements, user stories, constraints
   • •Extract: Technologies mentioned, patterns implied, performance requirements, integrations needed
3. •

   Detect project type from repository structure:

   • •Check for: frontend/, backend/, src/, app/, api/, ios/, android/, packages/
   • •Determine: Single project / Web app / Mobile app / Monorepo
   • •Set project type for planning decisions
4. •

   Fill Technical Context by analyzing spec:

   • •Language/Version: Detect from spec or mark NEEDS CLARIFICATION
   • •Primary Dependencies: Extract frameworks/libraries mentioned or mark NEEDS CLARIFICATION
   • •Storage: Identify database/storage needs or mark N/A
   • •Testing: Determine test frameworks needed or mark NEEDS CLARIFICATION
   • •Target Platform: Server/mobile/desktop/web or mark NEEDS CLARIFICATION
   • •Performance Goals: Extract from non-functional requirements or mark NEEDS CLARIFICATION
   • •Constraints: Identify technical constraints or mark NEEDS CLARIFICATION
   • •Scale/Scope: Estimate from requirements or mark NEEDS CLARIFICATION
5. •

   Identify research areas:

   • •For each NEEDS CLARIFICATION: Determine if external research needed
   • •For each unfamiliar technology/pattern: Flag for research
   • •For each architectural decision: Identify tradeoffs to research
   • •Generate research questions:
     • •"What are best practices for [technology] in [context]?"
     • •"How to choose between [option A] and [option B] for [use case]?"
     • •"What are performance implications of [approach]?"
6. •

   Save Technical Context and research needs to {STORY_DIR}/outline.md:

   markdown

   # Technical Outline: [STORY]
   
   **Date**: [YYYY-MM-DD]
   **Spec**: `spec.md`
   
   ## Project Type
   [Detected project type and structure]
   
   ## Technical Context
   [Filled technical context with NEEDS CLARIFICATION markers]
   
   ## Research Areas Identified
   
   ### High Priority (blocking decisions)
   1. **[Technology/Pattern]**: [Why research needed]
      - Question: [Specific research question]
      - Blocking: [What planning decision this blocks]
   
   ### Medium Priority (optimization decisions)
   [...]
   
   ### Low Priority (nice to have)
   [...]
   
   ## Suggested Research Queries
   - "[Generated research query 1]"
   - "[Generated research query 2]"
   
   Use the just-in-time research pattern (`.cursor/skills/just-in-time-research/SKILL.md`):
   - Do quick web search for each high-priority query
   - If web search is insufficient, generate a deep research prompt file and incorporate results
   
   ## Codebase Analysis Needed
   [Yes/No - if extending existing codebase]
   - Suggested: /story-scan --domain=[detected domain]
   
   ## Next Steps
   1. Use the suggested research queries (web search first; deep research prompt only if needed)
   2. Run /story-scan if extending existing codebase
   3. Run /story-plan to generate implementation plan (and embed/incorporate research)
   
   Or skip directly to /story-plan to use built-in just-in-time research (current behavior)
   

7. •

   Report completion:

   • •Path to outline.md
   • •Number of research areas identified (high/medium/low priority)
   • •Suggested next commands with specific research queries
   • •Note: "Run /story-plan directly to skip outline phase (uses built-in research)"

Behavior rules:

• •Mark things as NEEDS CLARIFICATION only if truly ambiguous from spec
• •Don't guess at technology choices - flag for research if unclear
• •Prioritize research areas by impact on architectural decisions
• •Generate specific, actionable research queries
• •If project structure unclear, mark for investigation but don't block

Context: $ARGUMENTS