lynx-trace-analysis

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

• Third-party content exposure detected (high risk: 0.80). The skill's trace-query workflow explicitly accepts a trace file path that "can be a URL" (see Trace Query Commands / Usage Examples in SKILL.md, e.g., --path "https://example.com/trace.pftrace"), so the agent will fetch and parse arbitrary external trace files (untrusted, user-provided) and must interpret their contents as part of analysis, which can materially influence its next actions.