add-brave

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The prompt instructs the agent to prompt the user for the BRAVE_API_KEY and then include that key verbatim in API request headers/commands to test connectivity, so the model must handle and output secret values directly, creating exfiltration risk.

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

• Third-party content exposure detected (high risk: 0.85). The required workflow calls the Brave Search API at runtime and then “summarizes and injects” the returned web results (titles/descriptions/snippets) into the agent’s LLM context, which is outsider-authored public web content fetched via an arbitrary search query.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.80). The skill performs runtime calls to the Brave Search API (e.g., https://api.search.brave.com/res/v1/web/search) and injects returned web results into the model context to drive agent responses, so external content fetched at runtime can directly influence prompts.