dspy-framework

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.80). The prompt contains multiple code examples that pass API keys and other secrets directly as string literals (e.g., api_key='YOUR_API_KEY', os.environ["LANGCHAIN_API_KEY"]="your-key"), which encourages embedding secrets verbatim in generated code/outputs and thus creates an exfiltration risk.

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

• Third-party content exposure detected (high risk: 0.80). The skill's SKILL.md explicitly shows the agent using retrieval and web search tools — e.g., dspy.Retrieve(...) and the ReAct example with tools=[web_search, calculator], plus ProductionRAG which retrieves passages and generates answers with citations — indicating it fetches and ingests externally sourced (potentially public/untrusted) web/passages that the agent reads and acts on.