busco-phylogeny

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

• Third-party content exposure detected (low risk: 0.10). In SKILL.md, the LLM executes WebFetch at runtime for “Substitution Model Recommendation” by retrieving IQ-TREE documentation from https://iqtree.github.io/doc/Substitution-Models, so outsider-authored text can be ingested if that external page is tampered with.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill includes runtime fetches of remote code/content that the agent will use or execute (e.g., WebFetch("https://iqtree.github.io/doc/Substitution-Models") and multiple wget commands that download and install/run Perl scripts and binaries from raw.githubusercontent.com / github.com such as https://raw.githubusercontent.com/PatrickKueck/FASconCAT-G/master/FASconCAT-G_v1.06.1.pl), which meet the criteria for external runtime dependencies that can control prompts or execute code.