rageprompt

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill requires placing the exact winning prompt text (pulled from local history) verbatim into a public "quote" field that will be uploaded, which could expose any secrets present in user prompts.

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

• Third-party content exposure detected (high risk: 0.85). Outsider-authored free text is ingested from the operating user’s local agent history files (e.g., .claude/projects/*.jsonl, .codex/sessions/*.jsonl, VS Code globalStorage state DBs, etc.) via scripts/collect-history.mjs, and then that text is fed into the “Judge” LLM context for scoring; these messages are not authored by the operating user (they are other-party chat prompts).