Steer towards frugal reads.

packages/core/src/prompts/snippets.ts

@@ -135,6 +135,7 @@ export function renderCoreMandates(options?: CoreMandatesOptions): string {
   return `
 # Core Mandates
 
+- **Context Efficiency:** Minimize context usage. Do not read entire files unless necessary. Use 'search_file_content' or 'read_file' with 'limit' to inspect large files.
 - **Conventions:** Rigorously adhere to existing project conventions when reading or modifying code. Analyze surrounding code, tests, and configuration first.
 - **Libraries/Frameworks:** NEVER assume a library/framework is available or appropriate. Verify its established usage within the project (check imports, configuration files like 'package.json', 'Cargo.toml', 'requirements.txt', 'build.gradle', etc., or observe neighboring files) before employing it.
 - **Style & Structure:** Mimic the style (formatting, naming), structure, framework choices, typing, and architectural patterns of existing code in the project.
@@ -370,7 +371,10 @@ function workflowStepUnderstand(options: PrimaryWorkflowsOptions): string {
     return `1. **Understand & Strategize:** Think about the user's request and the relevant codebase context. When the task involves **complex refactoring, codebase exploration or system-wide analysis**, your **first and primary action** must be to delegate to the 'codebase_investigator' agent using the 'codebase_investigator' tool. Use it to build a comprehensive understanding of the code, its structure, and dependencies. For **simple, targeted searches** (like finding a specific function name, file path, or variable declaration), you should use '${GREP_TOOL_NAME}' or '${GLOB_TOOL_NAME}' directly.`;
   }
   return `1. **Understand:** Think about the user's request and the relevant codebase context. Use '${GREP_TOOL_NAME}' and '${GLOB_TOOL_NAME}' search tools extensively (in parallel if independent) to understand file structures, existing code patterns, and conventions.
-Use '${READ_FILE_TOOL_NAME}' to understand context and validate any assumptions you may have. If you need to read multiple files, you should make multiple parallel calls to '${READ_FILE_TOOL_NAME}'.`;
+Use '${READ_FILE_TOOL_NAME}' to understand context and validate any assumptions you may have. If you need to read multiple files, you should make multiple parallel calls to '${READ_FILE_TOOL_NAME}'.
+Keep in mind the following as you explore:
+- Search and/or read enough files to gain a thorough understanding of the problem.
+- However, irrelevant context consumed by unnecessarily reading entire files will degrade the quality of your answer, so use ranged reads if you know which lines you need.`;
 }
 
 function workflowStepPlan(options: PrimaryWorkflowsOptions): string {

packages/core/src/tools/read-file.ts

@@ -177,12 +177,12 @@ export class ReadFileTool extends BaseDeclarativeTool<
           },
           offset: {
             description:
-              "Optional: For text files, the 0-based line number to start reading from. Requires 'limit' to be set. Use for paginating through large files.",
+              "Optional: For text files, the 0-based line number to start reading from. Requires 'limit' to be set. Use with 'limit' to target specific lines.",
             type: 'number',
           },
           limit: {
             description:
-              "Optional: For text files, maximum number of lines to read. Use with 'offset' to paginate through large files. If omitted, reads the entire file (if feasible, up to a default limit).",
+              "Optional: For text files, maximum number of lines to read. Use with 'offset' to paginate through large files. When checking for context or verifying changes, always set this to a small value (e.g. 50) to avoid reading the entire file.",
             type: 'number',
           },
         },