Bläddra i källkod

docs: add JSDoc documentation for ClineAsk and ClineSay types (#4427)

docs: add comprehensive JSDoc documentation for ClineAsk and ClineSay types
Hannes Rudolph 6 månader sedan
förälder
incheckning
91a477d831
1 ändrade filer med 54 tillägg och 0 borttagningar
  1. 54 0
      packages/types/src/message.ts

+ 54 - 0
packages/types/src/message.ts

@@ -4,6 +4,27 @@ import { z } from "zod"
  * ClineAsk
  */
 
+/**
+ * Array of possible ask types that the LLM can use to request user interaction or approval.
+ * These represent different scenarios where the assistant needs user input to proceed.
+ *
+ * @constant
+ * @readonly
+ *
+ * Ask type descriptions:
+ * - `followup`: LLM asks a clarifying question to gather more information needed to complete the task
+ * - `command`: Permission to execute a terminal/shell command
+ * - `command_output`: Permission to read the output from a previously executed command
+ * - `completion_result`: Task has been completed, awaiting user feedback or a new task
+ * - `tool`: Permission to use a tool for file operations (read, write, search, etc.)
+ * - `api_req_failed`: API request failed, asking user whether to retry
+ * - `resume_task`: Confirmation needed to resume a previously paused task
+ * - `resume_completed_task`: Confirmation needed to resume a task that was already marked as completed
+ * - `mistake_limit_reached`: Too many errors encountered, needs user guidance on how to proceed
+ * - `browser_action_launch`: Permission to open or interact with a browser
+ * - `use_mcp_server`: Permission to use Model Context Protocol (MCP) server functionality
+ * - `auto_approval_max_req_reached`: Auto-approval limit has been reached, manual approval required
+ */
 export const clineAsks = [
 	"followup",
 	"command",
@@ -27,6 +48,39 @@ export type ClineAsk = z.infer<typeof clineAskSchema>
  * ClineSay
  */
 
+/**
+ * Array of possible say types that represent different kinds of messages the assistant can send.
+ * These are used to categorize and handle various types of communication from the LLM to the user.
+ *
+ * @constant
+ * @readonly
+ *
+ * Say type descriptions:
+ * - `error`: General error message
+ * - `api_req_started`: Indicates an API request has been initiated
+ * - `api_req_finished`: Indicates an API request has completed successfully
+ * - `api_req_retried`: Indicates an API request is being retried after a failure
+ * - `api_req_retry_delayed`: Indicates an API request retry has been delayed
+ * - `api_req_deleted`: Indicates an API request has been deleted/cancelled
+ * - `text`: General text message or assistant response
+ * - `reasoning`: Assistant's reasoning or thought process (often hidden from user)
+ * - `completion_result`: Final result of task completion
+ * - `user_feedback`: Message containing user feedback
+ * - `user_feedback_diff`: Diff-formatted feedback from user showing requested changes
+ * - `command_output`: Output from an executed command
+ * - `shell_integration_warning`: Warning about shell integration issues or limitations
+ * - `browser_action`: Action performed in the browser
+ * - `browser_action_result`: Result of a browser action
+ * - `mcp_server_request_started`: MCP server request has been initiated
+ * - `mcp_server_response`: Response received from MCP server
+ * - `subtask_result`: Result of a completed subtask
+ * - `checkpoint_saved`: Indicates a checkpoint has been saved
+ * - `rooignore_error`: Error related to .rooignore file processing
+ * - `diff_error`: Error occurred while applying a diff/patch
+ * - `condense_context`: Context condensation/summarization has started
+ * - `condense_context_error`: Error occurred during context condensation
+ * - `codebase_search_result`: Results from searching the codebase
+ */
 export const clineSays = [
 	"error",
 	"api_req_started",