gr-core

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.80). The skill instructs the agent to read API keys from shared locations (Railway env, local files, and MEMORY.md which "has full key list"), forcing the LLM to access secret values in its context and creating a high risk those secrets could be output verbatim or leaked when invoking sub-skills.