Skip to content

Issue Classification Agent #180

Issue Classification Agent

Issue Classification Agent #180

Triggered via issue May 27, 2026 16:21
Status Success
Total duration 2m 51s
Artifacts 4
activation
13s
activation
detection
49s
detection
safe_outputs
11s
safe_outputs
call-handle-bug  /  activation
call-handle-bug / activation
call-handle-documentation  /  activation
call-handle-documentation / activation
call-handle-enhancement  /  activation
call-handle-enhancement / activation
call-handle-question  /  activation
call-handle-question / activation
call-handle-bug  /  agent
call-handle-bug / agent
call-handle-documentation  /  agent
call-handle-documentation / agent
call-handle-enhancement  /  agent
call-handle-enhancement / agent
call-handle-question  /  agent
call-handle-question / agent
call-handle-bug  /  detection
call-handle-bug / detection
call-handle-documentation  /  detection
call-handle-documentation / detection
call-handle-enhancement  /  detection
call-handle-enhancement / detection
call-handle-question  /  detection
call-handle-question / detection
call-handle-bug  /  safe_outputs
call-handle-bug / safe_outputs
call-handle-documentation  /  safe_outputs
call-handle-documentation / safe_outputs
call-handle-enhancement  /  safe_outputs
call-handle-enhancement / safe_outputs
call-handle-question  /  safe_outputs
call-handle-question / safe_outputs
call-handle-bug  /  conclusion
call-handle-bug / conclusion
call-handle-documentation  /  conclusion
call-handle-documentation / conclusion
call-handle-enhancement  /  conclusion
call-handle-enhancement / conclusion
call-handle-question  /  conclusion
call-handle-question / conclusion
conclusion
13s
conclusion
Fit to window
Zoom out
Zoom in

Annotations

2 errors and 3 warnings
detection
Reasons: Issue #1457 contained a prompt injection attempt disguised as a 'HackerOne authorized test' instructing the agent to post a probe-mark comment with tool names. The agent correctly detected and ignored the injection per security policy, taking no action (noop).
detection
🚨 Security threats detected: prompt injection
detection
⚠️ ERR_VALIDATION: ❌ Security threats detected: prompt injection Reasons: Issue #1457 contained a prompt injection attempt disguised as a 'HackerOne authorized test' instructing the agent to post a probe-mark comment with tool names. The agent correctly detected and ignored the injection per security policy, taking no action (noop).
safe_outputs
1 message(s) were cancelled because a code push operation failed
safe_outputs
🚫 Threat-detection warn policy aborted "noop" (Requirement WTD3): non-reviewable outputs must not be applied when detection conclusion is warning.

Artifacts

Produced during runtime
Name Size Digest
activation Expired
246 KB
sha256:64e6a40126142f33406cebeb748801f9cc40faa97630e5ef99ed4e858d35e4fd
agent
409 KB
sha256:ff68bc1c88fab6a7c1ebe582ef4fadf2a66dd6b22c3df92832e7b3bae3f61049
detection
3.39 KB
sha256:5c6c9f15bceeab0f95df8e0a9e10edaefe2a082deae1279ca2c5a10d01ecf88b
safe-outputs-items
301 Bytes
sha256:74aba260bdcb70d85f28d1e7cd09f9d0e3c2a39f0c0d060b8585c92a8c325105