rag-perf

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill explicitly logs the "fully resolved config as YAML" and the "shlex-joined aiperf command" to stdout (and notes the synthetic LLM endpoint may require auth), so any API keys or tokens placed in the config or command-line would be output verbatim and exposed.

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

• Third-party content exposure detected (medium risk: 0.30). The skill’s runtime can ingest outsider-authored free text when input.synthetic is used and the config points to an untrusted synthetic.prompts_file/synthetic.dataset_file (which then drives streamed LLM query generation).