skill-finder-cn

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

• Third-party content exposure detected (high risk: 0.80). The skill explicitly directs the agent to use the public PostPlus skills repository (e.g., the "Source Of Truth" commands in SKILL.md: npx -y skills add PostPlusAI/postplus-skills --global --list) and the provided scripts/search.sh invokes the same npx command to fetch listings from that public repo, meaning the agent will fetch and interpret untrusted, user-authored public repository content which can influence which skills are recommended and installed.

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

• Potentially malicious external URL detected (high risk: 0.90). I flagged the runtime npx fetch commands (e.g., "npx -y skills add PostPlusAI/postplus-skills --global --list" and "npx -y skills add PostPlusAI/postplus-skills --global --full-depth --skill '*' --agent ... --yes") because they fetch and install remote skill definitions at runtime which the agent relies on as the source of truth and which can directly control prompts/behavior.