get_file_contents
[analysis]
Read a bounded UTF-8 file window.
Unreleased read-only preview. Technical reference for agents and client integrators.
Before this call
Files must stay beneath the workspace without links. Text must be UTF-8 without NUL. An unindexed Java file requires coverage inspection and sync.
Parameters
| Name | Type | Required | Description |
|---|---|---|---|
| path | string | Yes | Contained workspace file path. Use a nonblank string of at most 4096 characters. |
| startLine | number | No | One-based integer line, from 1 to 2147483647. Default: 1 |
| lineCount | number | No | Integer from 1 to 1000. Default: 200 |
Examples
Technical examples for agents and client integrators. Replace paths and opaque keys with returned workspace values.
Illustrative arguments
Arguments
{
"path": "src/main/java/example/OrderService.java",
"startLine": 1,
"lineCount": 80
}Full MCP request
Full MCP request
{
"jsonrpc": "2.0",
"id": 1,
"method": "tools/call",
"params": {
"name": "get_file_contents",
"arguments": {
"path": "src/main/java/example/OrderService.java",
"startLine": 1,
"lineCount": 80
}
}
}Response notes
Reports text, contentVersion, span, readMode, windowTruncated, and outputTruncated. Java reads use snapshot text after a stale-file check. Other text reads use current_file. A UTF-8 BOM remains in text. Use smaller windows when output is truncated.