The Agent Skills Directory

[DATA_EXFILTRATION]: The skill communicates with the official SerpApi search endpoint (serpapi.com). Authentication is handled via API keys, and the instructions correctly recommend using environment variables (SERPAPI_API_KEY) to avoid hardcoding credentials.
[EXTERNAL_DOWNLOADS]: Provides instructions to download and install the serpapi-cli from the author's official GitHub repository and Homebrew tap. These are trusted vendor resources.
[PROMPT_INJECTION]: The skill ingests untrusted search result data from external web engines, representing a theoretical surface for indirect prompt injection.
Ingestion points: Search results (organic, news, images, etc.) are fetched from the serpapi.com API and processed by the agent.
Boundary markers: The skill does not define explicit delimiters to isolate search results from the agent's internal instructions.
Capability inventory: The skill does not contain scripts or instructions for performing high-risk actions such as arbitrary shell command execution, file system writes, or network requests to non-vendor domains.
Sanitization: The instructions recommend using mode="compact" to strip excessive metadata, which acts as a basic filter for the data processed by the LLM.

serpapi-search