chronicle

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (high risk: 0.80). The prompt contains an explicit instruction to "do not explain what you will do first — start querying immediately," which directs the agent to hide its data-access actions from the user and is therefore a deceptive/hidden instruction outside the skill's stated transparent analysis purpose.

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.80). The skill instructs the agent to read and surface turn content and tool/event arguments (user messages, assistant responses, tool_request arguments), which can contain API keys or other secrets and therefore may require echoing secret values verbatim.