referencer

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill instructs the agent to ask the user for their Referencer API key and to save it by embedding it verbatim into a command-line invocation (node "$RUNNER" init --api-key ...), which requires the LLM to handle/output the secret directly.

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

• Third-party content exposure detected (high risk: 0.75). Yes: the required query command reads free-form GraphQL text from either -f (readFileSync(fileFlag, "utf8")) or inline/stdin (args.positionals[0] / readStdin()), which is outsider-authored content at runtime if the operator points to a file/STDIN not authored by them.