match-requirement

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill embeds a literal API-style key ("sk-dev-proxy-123") in the Environment Variables section, meaning the LLM would see secret-like values in the prompt and could be asked or expected to reproduce them verbatim — an exfiltration risk.

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

• Third-party content exposure detected (medium risk: 0.30). In match_requirement.py, the “match” and “batch” commands POST attacker-controlled content indirectly by taking a user-supplied --chunk-key, sending it to the memory service (/match/requirement or /match/batch) where requirement_text is loaded and processed into obligation frames and evidence.