the cagent-output fence may not always be at the start of the response #197
derekmislertest.yml
on: pull_request
Prompt Sanitization Tests
5s
Output Extraction Tests
7s
Job Summary Format Tests
3s
Security Tests
7s
Exploit Tests
5s
Pirate Agent Test
2s
Invalid Agent Test
5s
Annotations
5 errors and 3 warnings
|
Invalid Agent Test
Process completed with exit code 1.
|
|
Invalid Agent Test
At least one API key is required. Provide one of: anthropic-api-key, openai-api-key, google-api-key, aws-bearer-token-bedrock, xai-api-key, nebius-api-key, or mistral-api-key
|
|
Prompt Sanitization Tests
═══════════════════════════════════════════════════════
|
|
Prompt Sanitization Tests
This strongly indicates a prompt injection attack
|
|
Prompt Sanitization Tests
🚨 HIGH-RISK pattern detected: show.*me.*(your|the|my).*(key|secret|token|api)
|
|
Prompt Sanitization Tests
Output will be scanned for actual secret leakage
|
|
Prompt Sanitization Tests
This PR modifies API key configuration - review carefully
|
|
Prompt Sanitization Tests
⚠️ MEDIUM-RISK pattern detected: ANTHROPIC_API_KEY
|