rc-codeprobe

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill pre-loads full source and config file contents and requires each finding's "evidence" to quote the relevant lines (and saves the report to disk), which will force the LLM to reproduce any hardcoded API keys, tokens, or passwords found in those files verbatim in its responses and saved report.

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

• Third-party content exposure detected (high risk: 0.85). The orchestrator reads the target project’s source files (which may include outsider-authored text) via Read and embeds their contents into the sub-skill agent prompts as === FILE: {filepath} === {content} === END FILE ===, creating an indirect prompt-injection path from arbitrary code/comments/strings in the reviewed repo.