lang-upgrade

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 1.00). The skill explicitly fetches and clones code and files from public GitHub (curl to https://raw.githubusercontent.com/... for VERSION in Step 3, git clone https://github.com/YiXinHui/langskill.git in Step 6, and reading README.md in Step 8) and then installs/replaces runtime skills from that untrusted public repository, which can directly modify agent behavior and thus enable indirect prompt injection.

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

• Potentially malicious external URL detected (high risk: 0.90). During runtime the script performs git clone https://github.com/YiXinHui/langskill.git (and also fetches https://raw.githubusercontent.com/YiXinHui/langskill/main/VERSION), and the cloned repository's skill directories are copied into ~/.claude/skills so remote code/skill content is fetched and installed—allowing that external content to control agent behavior.