Skip to content

Different approach to determining final confidence level of prompt injection evaluation outcomes #6729

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
dorien-koelemeijer merged 1 commit into from
Jan 29, 2026
Merged

Different approach to determining final confidence level of prompt injection evaluation outcomes #6729

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
67 changes: 34 additions & 33 deletions crates/goose/src/security/scanner.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -153,18 +153,29 @@ impl PromptInjectionScanner {
threshold
);

let final_result =
self.select_result_with_context_awareness(tool_result, context_result, threshold);
let final_confidence =
self.combine_confidences(tool_result.confidence, context_result.confidence);

tracing::info!(
"Security analysis complete: final_confidence={:.3}, malicious={}",
final_result.confidence,
final_result.confidence >= threshold
tool_confidence = %tool_result.confidence,
context_confidence = %context_result.confidence,
final_confidence = %final_confidence,
has_ml = tool_result.ml_confidence.is_some(),
has_patterns = !tool_result.pattern_matches.is_empty(),
threshold = %threshold,
malicious = final_confidence >= threshold,
"Security analysis complete"
);

let final_result = DetailedScanResult {
confidence: final_confidence,
pattern_matches: tool_result.pattern_matches,
ml_confidence: tool_result.ml_confidence,
};

Ok(ScanResult {
is_malicious: final_result.confidence >= threshold,
confidence: final_result.confidence,
is_malicious: final_confidence >= threshold,
confidence: final_confidence,
explanation: self.build_explanation(&final_result, threshold, &tool_content),
})
}
Expand Down Expand Up @@ -228,33 +239,23 @@ impl PromptInjectionScanner {
})
}

fn select_result_with_context_awareness(
&self,
tool_result: DetailedScanResult,
context_result: DetailedScanResult,
threshold: f32,
) -> DetailedScanResult {
let context_is_safe = context_result
.ml_confidence
.is_some_and(|conf| conf < threshold);

let tool_has_only_non_critical = !tool_result.pattern_matches.is_empty()
&& tool_result
.pattern_matches
.iter()
.all(|m| m.threat.risk_level != crate::security::patterns::RiskLevel::Critical);

if context_is_safe && tool_has_only_non_critical {
DetailedScanResult {
confidence: 0.0,
pattern_matches: Vec::new(),
ml_confidence: context_result.ml_confidence,
}
} else if tool_result.confidence >= context_result.confidence {
tool_result
} else {
context_result
fn combine_confidences(&self, tool_confidence: f32, context_confidence: f32) -> f32 {
// If tool is safe, context is not taken into account
if tool_confidence < 0.3 {
return tool_confidence;
}

if context_confidence < 0.3 {
return tool_confidence * 0.9;
}

if tool_confidence > 0.8 && context_confidence > 0.8 {
let max_conf = tool_confidence.max(context_confidence);
return (max_conf * 1.05).min(1.0);
}

// Default: weighted average (tool is primary signal)
tool_confidence * 0.8 + context_confidence * 0.2
}

async fn scan_with_classifier(
Expand Down
Loading