Documentation
¶
Index ¶
Constants ¶
View Source
const (
// DefaultMaxToolIterations is the default maximum number of tool execution loops.
DefaultMaxToolIterations = 25
)
Variables ¶
This section is empty.
Functions ¶
func ResultError ¶
func ResultError(result *formatter.ExecutionResult) error
ResultError converts an execution result into the error a command should return. It returns nil for a successful result that has something to show.
Errors that already explain themselves (user aborts, and provider errors that carry hints) are returned unchanged. Other failures are wrapped in ErrAIExecutionFailed. A successful result with no text and no tool activity is an error, so the command never prints a blank answer.
Types ¶
type Executor ¶
type Executor struct {
// contains filtered or unexported fields
}
Executor handles non-interactive AI execution with tool support.
func NewExecutor ¶
func NewExecutor(client ai.Client, toolExecutor *tools.Executor, atmosConfig *schema.AtmosConfiguration) *Executor
NewExecutor creates a new non-interactive executor.
type Options ¶
type Options struct {
// Prompt is the user's prompt/question.
Prompt string
// ToolsEnabled indicates if tool execution is allowed.
ToolsEnabled bool
// SessionID is an optional session ID for conversation context.
SessionID string
// IncludeContext includes stack context in the prompt.
IncludeContext bool
// History contains prior conversation messages to prepend before Prompt.
// Callers resuming a persisted session (see cmd/ai/session_helpers.go)
// populate this from the session's stored messages; it is empty for a
// fresh, session-less execution.
History []types.Message
}
Options contains execution options.
Source Files
Click to show internal directories.
Click to hide internal directories.