deslop

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.80). The skill tells the agent to "preserve meaning and useful detail" and to "keep facts, numbers, ... code, commands, paths" accurate, which can force the model to reproduce any secrets or API keys present in the source text verbatim and thus risks secret exfiltration.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). The SKILL.md defines an instruction to rewrite user-provided text (“default_prompt” in agents/openai.yaml says to “rewrite this”), so the required workflow ingests outsider-authored free text via the user’s input.