The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from Meta's Ad Library via the Apify platform. An attacker could craft a Facebook or Instagram ad containing malicious instructions that target the AI agent when the scraper fetches the ad creative.
Ingestion points: The agent receives external data through the output of scripts/search_meta_ads.py, particularly the ad_text, ad_creative_link_title, and ad_creative_link_description fields.
Boundary markers: The script does not wrap the scraped content in XML tags or other delimiters that would help the agent distinguish between data and instructions.
Capability inventory: While the script itself is limited, the agent consuming the output typically has broader capabilities such as file system access or shell execution, which could be exploited by injected instructions.
Sanitization: There is no filtering or sanitization performed on the ad text before it is returned to the agent.
[EXTERNAL_DOWNLOADS]: The skill relies on the Apify API (api.apify.com) to execute a web scraping actor. Apify is a well-known automation and scraping service.
[COMMAND_EXECUTION]: The skill's primary interface is a Python script (scripts/search_meta_ads.py) intended to be executed via the command line. It uses the standard argparse library to safely handle command-line arguments. The agent must ensure proper shell escaping when invoking this script with user-supplied inputs like company names or keywords.

meta-ad-scraper