The Agent Skills Directory

[Data Exposure & Exfiltration] (SAFE): The skill correctly uses the vm0_secrets manifest for the SCRAPENINJA_API_KEY. Network communication is restricted to legitimate API endpoints at rapidapi.com and apiroad.net. No unauthorized data access or exfiltration patterns were found.\n- [Command Execution] (SAFE): While the skill utilizes bash -c and curl, these are standard for making API requests and handling environment variables within the agent's environment. The usage is transparent and limited to the stated purpose.\n- [Indirect Prompt Injection] (LOW): As a scraping tool, this skill inherently processes untrusted data from the internet which could contain adversarial instructions.\n
Ingestion points: Content retrieved from scraped URLs in curl responses.\n
Boundary markers: None identified in the provided templates.\n
Capability inventory: Ability to write temporary files to /tmp, execute shell commands via bash -c, and perform network operations.\n
Sanitization: The skill does not perform sanitization of the scraped content before returning it to the agent.

scrapeninja