The Agent Skills Directory

[SAFE]: The skill implements tools (lightfast_search, lightfast_proxy_search, lightfast_proxy_call) designed to query and interact with various integrated service providers. These capabilities are consistent with the skill's stated purpose of facilitating team knowledge search and provider interaction.
[SAFE]: Untrusted user input is ingested via the $ARGUMENTS parameter and is wrapped in <user-request> boundary markers. This practice helps define the scope of user data for the agent, reducing the risk of indirect prompt injection. No other suspicious patterns such as obfuscation, unauthorized exfiltration, or remote code execution were detected.

lightfast-sdk