Skip to content

analysis-prompt.md

Latest commit

 

History

History
62 lines (49 loc) · 2.63 KB

analysis-prompt.md

File metadata and controls

62 lines (49 loc) · 2.63 KB
description Prompt for analyzing Claude Code best practices from engineering sources

Analyzing Best Methods for Using Claude Code

Analyze best practices for using Claude Code based on sources listed in @sources.csv (content available in source_content/ folder).

Process Flow

1. Filter & Categorize Content

  • Separate general vs. AI-specific practices: Identify content that applies to traditional software development vs. practices specifically important for agentic AI coding
  • Highlight fundamentals: Mark general engineering practices (version control, testing, CI) that have heightened importance when working with AI agents
  • Document criteria: Note why each practice is categorized as general or AI-specific

2. Organize & Identify Patterns

  • Group by semantic areas: Create logical categories (e.g., Testing Strategies, Prompt Engineering, Code Review, CI/CD Integration)
  • Highlight consensus: Identify practices mentioned in 2+ sources and note the frequency
  • Note contradictions: Flag conflicting recommendations with source attribution

3. Evaluate & Prioritize

Apply these criteria to each practice:

  • Impact: High/Medium/Low value for production software quality
  • Effort: Implementation complexity and learning curve
  • Specificity: How unique this is to Claude Code vs. general AI coding
  • Evidence: Whether the source provides data, examples, or anecdotal support

Flag items as:

  • ✅ Highly recommended (high impact, reasonable effort, well reasoned, well substantiated)
  • ⚠️ Context-dependent (valuable in specific scenarios)
  • ❌ Low priority (minimal impact or excessive effort or insufficient evidence provided)

Output Guidance

These are soft guidlines.

Format: Markdown report saved as claude-code-best-practices-report.md

Structure:

  1. Executive Summary (2-3 key takeaways)
  2. General Best Practices
    1. Best Practices that are eve more crucial in agentic AI coding workflows
  3. Core Recommendations by Category (with priority/importance/impact flags)
    1. Highlight consensus Findings (practices cited by multiple sources) in each category
    2. If sources conflict, present both views with source attribution rather than choosing
  4. Contradictions & Trade-offs
  5. Appendix:
    1. Source mapping table
    2. Complete set of recommendations

Style:

  • Target audience: Software engineers building production systems with Claude Code
  • Depth: Sufficient detail for actionability without excessive verbosity
  • Include code examples where relevant
  • Maximum length: 5000 words

Error Handling

  • If source_content/ is missing, STOP AND RAISE A FLAG.