spec-to-code-compliance

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). This skill requires quoting exact code and documentation excerpts (file + line numbers) and producing verbatim code excerpts/IR, which would force the model to reproduce any secrets (API keys, private keys, tokens, passwords) present in the provided materials.

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

• Third-party content exposure detected (high risk: 1.00). The skill explicitly accepts specification inputs as a file path or URL (commands/spec-compliance.md: "Spec document (PDF, MD, DOCX, HTML, TXT, or URL)") and lists WebFetch among allowed tools while Phase 0/Documentation Discovery requires fetching/open public docs (whitepapers, HTML, etc.), so the agent will fetch and interpret untrusted third‑party content as part of its workflow, which can materially influence subsequent actions.