What problem does it solve? AI coding agents often waste context window space by loading entire files when only a few lines are relevant. This Skill enforces a search-first discipline so agents find the right code with minimal file reads, keeping context clean for the actual task. ## Core Features & Use Cases - Search-First Workflow: A five-step process that defines the question, extracts keywords, searches with ripgrep/grep/Select-String, evaluates results, and reads only targeted line ranges. - Anti-Pattern Guardrails: Explicit examples of what to avoid, such as loading multiple full files or running overly broad searches. - Efficiency Metrics: Concrete benchmarks for files searched, files fully read, and lines read to measure context efficiency. - Use Case: Before fixing a bug in an unfamiliar repository, search for the failing function name, identify the two files that reference it, and read only the relevant 20-line sections instead of loading entire modules. ## Quick Start Ask the agent to find where user validation is implemented in the src directory using the context-fetch search-first approach.
