collaborating-with-gemini

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

• Attempt to modify system services in skill instructions detected (high risk: 0.80). The bridge explicitly allows Gemini to run shell commands and read/write workspace files by default (sandbox is opt-in) and even warns that approval-mode may not reliably prevent writes, so the skill can cause/modulate modifications to the host machine's state.

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

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