The Agent Skills Directory

[NO_CODE] (SAFE): The skill contains no executable Python or Node.js scripts. It consists entirely of markdown-based guidance for the AI agent.- [Indirect Prompt Injection] (SAFE): While the skill suggests gathering data from external sources such as GitHub repositories and API endpoints, it does not provide any code that would process this data in an unsafe manner or bypass standard agent guardrails.- [Prompt Injection] (SAFE): Analysis of the markdown content revealed no attempts to override system instructions, bypass safety filters, or extract system prompts.- [Data Exposure & Exfiltration] (SAFE): No hardcoded credentials, sensitive file paths, or network exfiltration patterns were detected in the instructions.

mteb-leaderboard