instantsearch

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The prompt instructs the agent to ask for the Algolia app ID and API key and to use them verbatim (e.g., in a curl command header), which requires the LLM to handle/output secret values directly.

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

• Third-party content exposure detected (high risk: 0.80). The skill's required Discover step explicitly tells the agent to fetch records from the user's Algolia index (see SKILL.md "Schema" with the curl query example), meaning it ingests third-party/user-provided data and uses that content to drive schema-driven rendering and routing decisions.