aris-auto-review-loop

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.80). The prompt allows the agent (in "nightmare" mode) to read the repository and other local config files and explicitly requires saving/pasting external reviewer outputs verbatim into project logs, which can cause any secrets found (API keys, tokens, or config contents) to be disclosed in LLM outputs and written files — an exfiltration risk.