prove-it

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill explicitly requires running commands/reading files and "report[ing] the raw output" and placing that raw output into the EVIDENCE field, which forces the agent to include any secrets present in those outputs verbatim.

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

• Third-party content exposure detected (medium risk: 0.30). The runtime path ingests the user-provided “argument to this skill” as the “claim to verify” and then uses WebSearch/Read/other tools based on that free text, so outsider-authored text can be submitted into the workflow’s verification/research steps.