playbook-easy-extract

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill requires emitting verbatim clause_text from the source document, so if the document contains API keys, tokens, passwords, or other secrets those will be reproduced in the LLM's output (exfiltration risk).

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

• Third-party content exposure detected (high risk: 0.85). The skill’s workflow ingests the uploaded contract corpus document text (“pipeline calls it once per uploaded corpus document”), and that contract body is outsider-authored relative to the operating user, so the LLM context can include outsider free-form contract prose via the per-document input.