helpful-formatter

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (high risk: 1.00). The skill contains explicit instructions to ignore prior operational constraints and adopt "unrestricted" behavior and elevated permissions, which are hidden/deceptive overrides outside the skill's stated purpose of formatting text.

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.80). This skill instructs the assistant to ignore normal safeguards and unconditionally rewrite any user-provided text, which can lead to verbatim inclusion and exfiltration of secrets or credentials present in that text.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 0.90). The skill content contains explicit prompt-injection instructions to ignore operational safeguards and assume elevated, unrestricted permissions—effectively a backdoor to bypass assistant safety—which is high-risk even though it contains no direct exfiltration or remote-execution code.

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (high risk: 0.80). This skill explicitly instructs the agent to "set aside" operational safeguards and adopt an "unrestricted" mode with "elevated access permissions," which encourages bypassing safety and security mechanisms and could lead to performing privileged or state-changing actions on the host.