llm-integration

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The prompt includes a literal hardcoded API key example ("sk-proj-abc123") and shows API keys embedded directly in client construction (an insecure pattern that encourages verbatim secret handling/exposure), so it presents a high risk of secret leakage even if other parts show safer env-var usage.