The Agent Skills Directory

[DATA_EXFILTRATION]: The skill transmits user-provided URLs and search parameters to Bright Data's official API endpoints (api.brightdata.com). This behavior is the intended primary purpose of the skill and is directed to the vendor's own infrastructure.
[COMMAND_EXECUTION]: Shell scripts (datasets.sh and fetch.sh) are used to orchestrate the data fetching process. The implementation utilizes jq to handle user input, ensuring that data is properly escaped before being included in API requests.
[PROMPT_INJECTION]: The skill processes data from external websites, creating a surface for indirect prompt injection.
Ingestion points: Data is retrieved from third-party websites (Amazon, LinkedIn, etc.) via the Bright Data API and returned to the agent context.
Boundary markers: The output is returned as structured JSON, but no explicit boundary markers or instructions to ignore embedded commands are included in the prompt interpolation.
Capability inventory: The environment has access to network tools (curl) and shell execution (bash).
Sanitization: The skill returns the raw JSON response from the API without additional filtering or sanitization of the scraped content.

data-feeds