ax

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The prompt includes examples and tooling that expect Authorization headers or -u credentials to be placed inline in generated commands (e.g., -H 'authorization: Bearer x'), which encourages embedding real API tokens/passwords verbatim in LLM outputs and thus creates an exfiltration risk.

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

• Third-party content exposure detected (medium risk: 0.30). The skill’s runtime workflow explicitly performs HTTP fetches/extraction from a user-supplied URL (e.g., ax <url> --outline/--locate/--row), which means outsider-authored free text from arbitrary web pages or API responses can be ingested by the LLM during extraction.