agentfortress 1.0.0

Runtime protection and security monitoring for AI agents — The CrowdStrike for AI Agents
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43

use crate::models::{PolicyAction, PolicyActionKind};

pub struct Scanner;

impl Scanner {
    pub fn new() -> Self {
        Self
    }

    pub fn scan(&self, text: &str) -> PolicyAction {
        let text_lower = text.to_lowercase();
        let injection_patterns = [
            "ignore previous instructions",
            "ignore all instructions",
            "ignore above instructions",
            "you are now",
            "disregard your system",
            "disregard your previous",
            "forget everything",
            "forget all",
            "jailbreak",
            "dan mode",
        ];
        for pattern in &injection_patterns {
            if text_lower.contains(pattern) {
                return PolicyAction {
                    action: PolicyActionKind::Block,
                    reason: Some(format!("Prompt injection pattern detected: '{}'", pattern)),
                };
            }
        }
        PolicyAction {
            action: PolicyActionKind::Allow,
            reason: None,
        }
    }
}

impl Default for Scanner {
    fn default() -> Self {
        Self::new()
    }
}