conversation-memory

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill explicitly instructs the LLM to extract and store "facts" from user messages (via llm.complete on Message: "${message.content}") and later includes those facts verbatim in prompts/responses, which would cause any embedded secrets (API keys, passwords, tokens) to be reproduced and persisted in outputs and memory.