⚡ Tool: context

The context tool is the interface between the AI agent and your project's Language Server Protocol (LSP). It provides 100% accurate, compiler-verified code definitions.

Technical Implementation

Standard Zed utilizes a background regex-based semantic indexer. This fork disables that subsystem and replaces it with crates/agent/src/tools/context_tool.rs.

When the agent needs to understand a symbol, it forwards the request directly to the active server (rust-analyzer, pyright, etc.). This ensures the agent is reading the actual resolved AST of your project, not just a text-based guess.

Usage Examples

Example 1: Locating Type Definitions

User

Where is the `spawn` method defined in the GPUI crate?

Zed Agent

Querying LSP for `spawn` definition...

Tool Output

@context({ query: "spawn" }) >> Found: crates/gpui/src/executor.rs:L112

Example 2: Understanding Trait Implementations

User

Show me all implementations of the `Render` trait for components in `crates/ui`.

Zed Agent

Requesting LSP implementation map for `Render` trait...

Tool Output

@context({ query: "Render" }) >> Impl 1: crates/ui/src/button.rs >> Impl 2: crates/ui/src/modal.rs

Workflow Impact & Impact

• Zero Hallucination: The agent sees exactly what the compiler sees. It never guesses a file path or a symbol location.
• Contextual Precision: Understands the difference between model.save() and result.save() even if they share the same method name.
• Zero-Config Performance: Works instantly with your existing LSP setup (rust-analyzer, gopls, etc.) without needing a background index.
• Massive Scale: Can find a needle in a haystack across 1M+ lines of code by offloading the heavy lifting to the Language Server.