The Agent Skills Directory

[SAFE]: The skill is entirely descriptive and instructional, designed to elevate an agent's reasoning capabilities using established mental models from scientific literature and public discourse. No malicious patterns or security risks were identified across the 77 files.- [NO_CODE]: The skill does not package any executable code, scripts, or binaries. It functions purely through prompt-based personality steering and reference data.- [EXTERNAL_DOWNLOADS]: The skill includes extensive citations to external sources including the Nobel Prize organization, Google DeepMind, and reputable news outlets (The New York Times, Wired, Lex Fridman). These are provided for attribution and research context; no automated remote code execution or suspicious network fetching patterns are present.

demis-hassabis