Struct UserPromptSubmitEvent 

pub struct UserPromptSubmitEvent {
    pub prompt: String,
    pub history: Vec<Value>,
}

Event fired before processing user input, enabling content moderation and prompt enhancement.

This event is triggered whenever a user submits a prompt to the agent, before the agent begins processing it. Use this to implement content moderation, add context, inject instructions, or track user interactions.

Use Cases

• Content moderation: Filter inappropriate or harmful user inputs
• Prompt enhancement: Add system context, timestamps, or user information
• Input validation: Ensure prompts meet format or length requirements
• Usage tracking: Log user interactions for analytics or billing
• Context injection: Add relevant background information to every prompt

Fields

• prompt: The user’s original input text
• history: Read-only snapshot of the conversation history before this prompt

Example: Content Moderation

use open_agent::{UserPromptSubmitEvent, HookDecision};

async fn content_moderator(event: UserPromptSubmitEvent) -> Option<HookDecision> {
    // Block prompts containing banned words
    let banned_words = ["spam", "malware", "hack"];

    for word in banned_words {
        if event.prompt.to_lowercase().contains(word) {
            return Some(HookDecision::block(
                format!("Content policy violation: contains '{}'", word)
            ));
        }
    }
    None // Allow clean prompts
}

Example: Automatic Context Enhancement

use open_agent::{UserPromptSubmitEvent, HookDecision};

async fn add_context(event: UserPromptSubmitEvent) -> Option<HookDecision> {
    // Add helpful context to every user prompt
    let enhanced = format!(
        "{}\n\n---\nContext: User timezone is UTC, current session started at 2025-11-07",
        event.prompt
    );

    Some(HookDecision::modify_prompt(
        enhanced,
        "Added session context"
    ))
}

Example: Usage Tracking

use open_agent::{UserPromptSubmitEvent, HookDecision};

async fn track_usage(event: UserPromptSubmitEvent) -> Option<HookDecision> {
    // Log every user interaction for analytics
    println!(
        "[ANALYTICS] User submitted prompt of {} characters at history depth {}",
        event.prompt.len(),
        event.history.len()
    );

    // Could also:
    // - Update usage quotas
    // - Send to analytics service
    // - Check rate limits

    None // Don't modify the prompt
}

Modification Behavior

If you return HookDecision::modify_prompt(), the modified prompt completely replaces the original user input before the agent processes it. This is powerful but should be used carefully to avoid confusing the user or the agent.

Fields

prompt: String

The user’s original input prompt text

history: Vec<Value>

Snapshot of conversation history (read-only) - does not include this prompt yet

Implementations

impl UserPromptSubmitEvent

pub fn new(prompt: String, history: Vec<Value>) -> Self

Creates a new UserPromptSubmitEvent.

This constructor is typically called by the agent runtime when processing user input, not by user code. Users receive instances of this struct in their hook handlers.

Trait Implementations

impl Clone for UserPromptSubmitEvent

fn clone(&self) -> UserPromptSubmitEvent

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for UserPromptSubmitEvent

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.