stock-earnings-review

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

• Third-party content exposure detected (high risk: 0.75). Outsider free text can enter the LLM context via the runtime API response field data.content (and data.title) from scripts/call_review_api.py → extract_comment_response_fields() → returned content/title, which is then used for the agent’s final response; this content is not authored by the operating user.

HIGH W008: Secret detected in skill content (API keys, tokens, passwords).

• Secret detected (high risk: 1.00). I scanned the code and found a hardcoded, high-entropy API key string in scripts/common.py:

EM_API_KEY = os.environ.get("EM_API_KEY", "em_fjFqd4YB6Cqs52LF48XWbMDdLNq6MyNg").strip() ...

This is a literal default value that looks like a real API key (high-entropy, not a placeholder). It is used in auth_headers/base_headers, so it is a usable credential if not overridden by an environment variable. Other literals (API endpoint URLs, directory paths, and the placeholder constant _API_KEY_PLACEHOLDER) are not secrets and are intentionally ignored.