pathogen-variant-surveillance

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

• Third-party content exposure detected (high risk: 0.75). The required runtime workflow fetches and parses outsider-authored free text from public sources (e.g., https://raw.githubusercontent.com/.../alias_key.json and lineage_notes.txt, plus remote /sample/databaseConfig and error detail bodies from user-supplied --base-url), and then emits those strings into agent-readable table/TSV/JSON output.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill fetches pango-designation files at runtime from https://raw.githubusercontent.com/cov-lineages/pango-designation/master/pango_designation/alias_key.json and https://raw.githubusercontent.com/cov-lineages/pango-designation/master/lineage_notes.txt which are parsed and injected into agent-visible output (used to unalias lineages and mark withdrawals), so remote content directly controls runtime behavior and what the agent reports.