Skip to content

Update prompt injection detection metrics (due to errors exporting to datadog) #5692

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 2 commits into from
Nov 13, 2025
Merged

Update prompt injection detection metrics (due to errors exporting to datadog) #5692

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
db681df
update prompt injection metrics due to errors exporting to datadog
dorien-koelemeijer Nov 12, 2025
3406daf
remove emojis from traces
dorien-koelemeijer Nov 13, 2025
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
5 changes: 2 additions & 3 deletions crates/goose/src/agents/tool_execution.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -87,9 +87,8 @@ impl Agent {
tracing::info!(
counter.goose.prompt_injection_user_decisions = 1,
decision = ?confirmation.permission,
"🔒 User security decision: {:?} for finding ID: {}",
confirmation.permission,
finding_id
finding_id = %finding_id,
"User security decision"
);
}

Expand Down
22 changes: 11 additions & 11 deletions crates/goose/src/security/mod.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -48,16 +48,16 @@ impl SecurityManager {
) -> Result<Vec<SecurityResult>> {
if !self.is_prompt_injection_detection_enabled() {
tracing::debug!(
gauge.goose.prompt_injection_scanner_enabled = 0,
"🔓 Security scanning disabled"
counter.goose.prompt_injection_scanner_disabled = 1,
"Security scanning disabled"
);
return Ok(vec![]);
}

let scanner = self.scanner.get_or_init(|| {
tracing::info!(
gauge.goose.prompt_injection_scanner_enabled = 1,
"🔓 Security scanner initialized and enabled"
counter.goose.prompt_injection_scanner_enabled = 1,
"Security scanner initialized and enabled"
);
PromptInjectionScanner::new()
});
Expand All @@ -79,26 +79,26 @@ impl SecurityManager {

// Get threshold from config - only flag things above threshold
let config_threshold = scanner.get_threshold_from_config();
let sanitized_explanation = analysis_result.explanation.replace('\n', " | ");

if analysis_result.is_malicious {
let above_threshold = analysis_result.confidence > config_threshold;
let finding_id = format!("SEC-{}", Uuid::new_v4().simple());

tracing::warn!(
counter.goose.prompt_injection_finding = 1,
gauge.goose.prompt_injection_confidence_score = analysis_result.confidence,
above_threshold = above_threshold,
tool_name = %tool_call.name,
tool_request_id = %tool_request.id,
confidence = analysis_result.confidence,
explanation = %analysis_result.explanation,
explanation = %sanitized_explanation,
finding_id = %finding_id,
threshold = config_threshold,
"{}",
if above_threshold {
"🔒 Current tool call flagged as malicious after security analysis (above threshold)"
"Current tool call flagged as malicious after security analysis (above threshold)"
} else {
"🔒 Security finding below threshold - logged but not blocking execution"
"Security finding below threshold - logged but not blocking execution"
}
);
if above_threshold {
Expand All @@ -116,7 +116,7 @@ impl SecurityManager {
tool_name = %tool_call.name,
tool_request_id = %tool_request.id,
confidence = analysis_result.confidence,
explanation = %analysis_result.explanation,
explanation = %sanitized_explanation,
"✅ Current tool call passed security analysis"
);
}
Expand All @@ -125,8 +125,8 @@ impl SecurityManager {

tracing::info!(
counter.goose.prompt_injection_analysis_performed = 1,
"🔍 Security analysis complete - found {} security issues in current tool requests",
results.len()
security_issues_found = results.len(),
"Security analysis complete"
);
Ok(results)
}
Expand Down
Loading