The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes local bash scripts (scripts/scrape.sh and scripts/search.sh) that execute curl and jq to interface with external APIs and manage data securely.
[EXTERNAL_DOWNLOADS]: The skill initiates network requests to api.brightdata.com, a well-known service provider for proxy and scraping solutions, to retrieve data.
[PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it processes data from untrusted external websites. 1. Ingestion points: Data is ingested from arbitrary URLs via the scraping and search scripts. 2. Boundary markers: The scripts do not implement specific delimiters or 'ignore' instructions for the content returned from the API. 3. Capability inventory: The skill provides the agent with capabilities to perform authenticated network requests and data processing. 4. Sanitization: No content sanitization or validation is performed on the scraped markdown or search descriptions before they are returned to the agent.

brightdata