friction-log

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill mandates writing the user's initial request verbatim into the friction log (and then printing the log to stdout), which forces any secrets present in the user's prompt to be output exactly as provided.

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

• Third-party content exposure detected (high risk: 0.75). The required workflow explicitly instructs the agent to read “any URLs in the prompt” and fetch their contents before starting (SKILL.md:14), which can include outsider-authored free text from public web pages and thus be injected into the LLM context.