A
92/100 First Seen
Feb 18, 2026
Last Scanned
Feb 22, 2026
Findings
2
Score
92/100
MEDIUM 1
LOW 1
Findings (2)
MEDIUM
Base64-encoded instructions
L135 Detects instructions to decode and execute base64 content
decoder_decode","arguments":{"token":"eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9 FIX
Remove text that simulates multi-turn conversations or fake user/assistant message pairs. These patterns attempt to manipulate the agent by fabricating conversation history.
FP?
Likely FP if the matched text is documentation showing example conversations or API request/response formats for illustration purposes.
LOW
JWT token
L135 Detects JWT tokens
eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJzdWIiOiIxMjM0NTY3ODkwIiwibmFtZSI6IkpvaG4gRG9lIiwiaWF0IjoxNTE2MjM5MDIyfQ.SflKxwRJSMeKKF2QT4fwpMeJf36POk6yJV_adQssw5c FIX
Remove hardcoded JWT tokens from the skill definition. Generate tokens dynamically at runtime and set appropriate expiration times.
FP?
Likely FP if the matched text is a documentation example showing JWT structure with clearly fake values, or an expired demo token.