fable-safe-prompt

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill explicitly requires returning the user's prompt "in full, verbatim" (operating on the text inside ...) and instructs copying that full prompt to the clipboard, so any API keys/secrets included by the user would be echoed verbatim in the agent's output and exposed.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). This skill explicitly instructs how to surgically rewrite user prompts to evade model safety classifiers (including removing "show your reasoning") and even suggests fallback routes, enabling intentional safety-filter circumvention.