serena-code-reading

What problem does it solve?

Reading entire code files for context is token-inefficient and slow for AI agents. This Skill enforces Serena's progressive disclosure pattern, achieving 90-95% token savings by reading only what's necessary.

Core Features & Use Cases

• Progressive Disclosure Protocol: Mandates starting with a file overview, then checking symbol signatures, and finally reading specific symbol bodies, saving thousands of tokens per operation.
• Core Tools Reference: Provides guidance on get_symbols_overview (file outline), find_symbol (specific symbol details), replace_symbol_body (precise editing), and find_referencing_symbols (impact analysis).
• Token Impact Comparison: Clearly demonstrates how Serena's approach reduces token usage from ~5,000 tokens for a full file read to ~350 tokens for a targeted exploration.
• Use Case: Instead of reading a 2,000-line file (5,000 tokens), use Serena to get an overview, check a signature, and read a specific method body for a total of ~350 tokens.

Quick Start

Get an overview of the file 'src/services/auth.service.ts' to understand its structure.