frontend-beetl

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

• Third-party content exposure detected (high risk: 0.75). SKILL.md:7-10 describes runtime fetching/caching of documentation content (via python3 scripts/cache_docs.py ... --permalink <permalink> and reading __00.md/__01.md~__NN.md), which is likely sourced from external web documentation pages (outsider-authored free text) and then ingested into the LLM context for answering.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly runs "python3 scripts/cache_docs.py --skill frontend-beetl --permalink " at runtime to fetch and cache documentation from external URLs such as https://jeesite.com/docs/views-beetl/, and those fetched documents are read and used to construct answers, so the remote content directly controls the agent's prompts/context.