vtcode_commons/
sanitizer.rs

1//! Secret sanitization utilities for redacting sensitive information.
2//!
3//! Provides regex-based secret redaction for:
4//! - OpenAI API keys (`sk-...`)
5//! - AWS Access Key IDs (`AKIA...`)
6//! - Bearer tokens (`Bearer ...`)
7//! - Generic secret assignments (`api_key=...`, `password:...`, etc.)
8//!
9//! Use this module to sanitize text before logging, displaying in UI,
10//! or storing in session archives.
11
12use regex::Regex;
13use std::sync::LazyLock;
14
15/// OpenAI API key pattern: sk- followed by alphanumeric characters
16static OPENAI_KEY_REGEX: LazyLock<Regex> = LazyLock::new(|| compile_regex(r"sk-[A-Za-z0-9]{20,}"));
17
18/// AWS Access Key ID pattern: AKIA followed by 16 alphanumeric characters
19static AWS_ACCESS_KEY_ID_REGEX: LazyLock<Regex> = LazyLock::new(|| compile_regex(r"\bAKIA[0-9A-Z]{16}\b"));
20
21/// Bearer token pattern: "Bearer " followed by token characters
22static BEARER_TOKEN_REGEX: LazyLock<Regex> = LazyLock::new(|| compile_regex(r"(?i)\bBearer\s+[A-Za-z0-9.\-_]{16,}\b"));
23
24/// Generic secret assignment pattern: key=value or key: value format
25/// Matches common secret key names like api_key, token, secret, password
26static SECRET_ASSIGNMENT_REGEX: LazyLock<Regex> =
27    LazyLock::new(|| compile_regex(r#"(?i)\b(api[\-_]?key|token|secret|password)\b(\s*[:=]\s*)(["']?)[^\s"']{8,}"#));
28
29/// Redact secrets and sensitive keys from a string.
30///
31/// This is a best-effort operation using well-known regex patterns.
32/// Redacted values are replaced with `[REDACTED_SECRET]`.
33///
34/// # Examples
35///
36/// ```
37/// use vtcode_commons::sanitizer::redact_secrets;
38///
39/// let input = "Found key: sk-test1234567890abcdefghij".to_string();
40/// let output = redact_secrets(input);
41/// assert_eq!(output, "Found key: [REDACTED_SECRET]");
42/// ```
43pub fn redact_secrets(input: String) -> String {
44    let r1 = OPENAI_KEY_REGEX.replace_all(&input, "[REDACTED_SECRET]");
45    let r2 = AWS_ACCESS_KEY_ID_REGEX.replace_all(&r1, "[REDACTED_SECRET]");
46    let r3 = BEARER_TOKEN_REGEX.replace_all(&r2, "Bearer [REDACTED_SECRET]");
47    let r4 = SECRET_ASSIGNMENT_REGEX.replace_all(&r3, "$1$2$3[REDACTED_SECRET]");
48    // `into_owned` clones only when the final result is `Borrowed` (no regex
49    // matched at all); when any redaction occurred it moves the owned string
50    // without an extra allocation. Do NOT short-circuit on `Cow::Borrowed` —
51    // the final Cow is `Borrowed` whenever the *last* regex doesn't match,
52    // even if earlier regexes did, which would silently discard redactions.
53    r4.into_owned()
54}
55
56#[allow(clippy::panic)]
57fn compile_regex(pattern: &str) -> Regex {
58    match Regex::new(pattern) {
59        Ok(regex) => regex,
60        // Panic is acceptable thanks to the `load_regex` test
61        Err(err) => panic!("invalid regex pattern `{pattern}`: {err}"),
62    }
63}
64
65#[cfg(test)]
66mod tests {
67    use super::*;
68
69    #[test]
70    fn load_regex() {
71        // Verify all regex patterns compile without panicking
72        let _ = redact_secrets("test".to_string());
73    }
74
75    #[test]
76    fn redacts_openai_key() {
77        let input = "Found key: sk-test1234567890abcdefghij".to_string();
78        let output = redact_secrets(input);
79        assert_eq!(output, "Found key: [REDACTED_SECRET]");
80    }
81
82    #[test]
83    fn redacts_aws_access_key() {
84        // AKIAIOSFODNN7EXAMPLE is AWS's well-known documentation example key.
85        let input = " creds: AKIAIOSFODNN7EXAMPLE ".to_string();
86        let output = redact_secrets(input);
87        assert_eq!(output, " creds: [REDACTED_SECRET] ");
88    }
89
90    #[test]
91    fn redacts_bearer_token() {
92        let input = "Authorization: Bearer eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9".to_string();
93        let output = redact_secrets(input);
94        assert_eq!(output, "Authorization: Bearer [REDACTED_SECRET]");
95    }
96
97    #[test]
98    fn redacts_api_key_assignment() {
99        let input = "api_key=sk-test12345678".to_string();
100        let output = redact_secrets(input);
101        assert_eq!(output, "api_key=[REDACTED_SECRET]");
102    }
103
104    #[test]
105    fn redacts_password_assignment() {
106        let input = "password: mysecretvalue".to_string();
107        let output = redact_secrets(input);
108        assert_eq!(output, "password: [REDACTED_SECRET]");
109    }
110
111    #[test]
112    fn redacts_token_in_quotes() {
113        let input = r#"token="abc123xyz789abcdef""#.to_string();
114        let output = redact_secrets(input);
115        assert_eq!(output, r#"token="[REDACTED_SECRET]""#);
116    }
117
118    #[test]
119    fn preserves_short_values() {
120        // Values under 8 characters should not be redacted
121        let input = "password: short".to_string();
122        let output = redact_secrets(input);
123        assert_eq!(output, "password: short");
124    }
125
126    #[test]
127    fn redacts_multiple_secrets() {
128        let input = "Keys: sk-test1234567890abcdefghij and AKIAIOSFODNN7EXAMPLE".to_string();
129        let output = redact_secrets(input);
130        // Verify both secrets are redacted
131        assert!(output.contains("[REDACTED_SECRET]"));
132        assert!(!output.contains("AKIAIOSFODNN7EXAMPLE"));
133        assert!(!output.contains("sk-test1234567890abcdefghij"));
134    }
135
136    #[test]
137    fn preserves_non_secret_text() {
138        let input = "Hello world, this is normal text".to_string();
139        let output = redact_secrets(input);
140        assert_eq!(output, "Hello world, this is normal text");
141    }
142}
vtcode_commons/sanitizer.rs

vtcode_commons/
sanitizer.rs
