secret-scanning

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill explicitly instructs the agent to take user- or file-provided content (which may contain secrets) and pass it verbatim to the run_secret_scanning tool / MCP (and shows examples with actual secret strings), so the LLM must handle and emit secret values in tool inputs/outputs.

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

• Third-party content exposure detected (low risk: 0.10). The skill only calls GitHub’s run_secret_scanning on content explicitly provided by the user as a code snippet, specific file (via the user-directed Read tool), or staged diff, rather than ingesting any outsider-authored free text from a queue/feed without selecting it first.