boris

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly instructs a runtime WebFetch to https://howborisusesclaudecode.com/api/version and to offer running curl to download SKILL.md from https://howborisusesclaudecode.com/api/install, which means remote content would be fetched at runtime and could replace/control the skill's prompts/instructions.

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

• Attempt to modify system services in skill instructions detected (high risk: 0.80). The skill explicitly instructs users/agents to disable or bypass permission prompts (e.g. using "--permission-mode=dontAsk" or "--dangerously-skip-permissions"), which encourages bypassing security mechanisms on the host.