linkdrop-x-0x-kaize-awesome-claude-code-repos

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

• Third-party content exposure detected (high risk: 0.85). The skill’s runtime workflow fetches and parses outsider-authored public web content from https://github.com/hesreallyhim/awesome-claude-code (via gh api .../readme | base64 -d > /tmp/acc-readme.md and subsequent grep), which becomes readable free text in the agent’s context for candidate selection.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly fetches the live README at runtime (gh api repos/hesreallyhim/awesome-claude-code/readme -> https://github.com/hesreallyhim/awesome-claude-code) and instructs cloning repos (gh repo clone /), so external README content and cloned repo files are pulled into the agent's runtime context and can directly influence recommendations/prompts or contain executable hooks/scripts — a required runtime dependency.