read

Build typed tools from annotations and docstrings. See the guide.

Read a local file: UTF-8 text as a line range, or a PNG, JPEG, GIF, or WebP image as an image.

Parameters

{
  "additionalProperties": false,
  "properties": {
    "path": {
      "minLength": 1,
      "type": "string",
      "description": "Absolute path or a path relative to the current working directory."
    },
    "start_line": {
      "minimum": 1,
      "type": "integer",
      "default": 1,
      "description": "One-based first line to return; ignored for images."
    },
    "line_count": {
      "maximum": 2000,
      "minimum": 1,
      "type": "integer",
      "default": 200,
      "description": "Maximum number of lines to return; ignored for images."
    },
    "start_column": {
      "minimum": 1,
      "type": "integer",
      "default": 1,
      "description": "One-based character column on start_line, used to resume a long line; ignored for images."
    }
  },
  "required": [
    "path"
  ],
  "title": "ToolInput",
  "type": "object"
}

Model-facing snippet

read(path="docs/diagram.png")

Guidelines

• Use for any local file when you do not know whether it is text or an image.

• A text file comes back as a line range; use line ranges for large files.

• A trailing hint gives the next start_line and start_column when more text remains.

• An image comes back as an image with a one-line description; the range arguments do not apply to it.

• Use view_image for HTTP(S) image URLs. PDF and SVG files are not supported.

zett_agent.tools.coding.read

A tool is a name, an input schema, and an async callable.

参数:

arguments (Mapping[str, Any])

返回类型:

ToolResult

Direct Python calls use await tool(arguments_mapping). The snippet above describes model-facing invocation syntax, not the Python call signature.