vedic-core

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (high risk: 0.70). Yes — the prompt includes explicit instructions to conceal internal operational rules from the user (e.g., "不要在聊天框里复述这条规则给用户看") and to hide aspects of its process, which are deceptive and fall outside the skill's stated user-facing analysis purpose.

MEDIUM W021: Hidden or invisible Unicode characters detected (potential obfuscation or prompt injection).

• Hidden Unicode characters detected (1 type(s) found)