nuke-verify

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill mandates capturing and embedding verbatim file lines and gate outputs into findings and verify.md, which forces an agent to reproduce any secrets present in those files or outputs verbatim (exfiltration risk).

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

• Third-party content exposure detected (medium risk: 0.30). In Nuke Verify, the runtime explicitly ingests user/outsider-authored free text when [intent] is provided as free text (Phase 0 preflight step “Resolve the intent source … explicit intent: file or text argument” which is then distilled into the numbered requirement list and used throughout validation/fix).