Ranked 19th on GitHub's all-time Star leaderboard (surpassing 200K Stars) and with over 13.5 million downloads, Skills has officially released the major v1.2 update. This upgrade marks SKILL.md's evolution from a single-point practice in the Claude community to a cross-platform, standardized new paradigm for software engineering: Not only has it been listed on the official Claude plugin marketplace, but it also adds full support for OpenAI Codex via agents/openai.yaml, officially enabling cross-platform invocation of Agent skills. High-frequency Skill experience revamp: /grilling (Requirement Interrogation): Upgraded from tedious "one-question-at-a-time" to "multi-round focused questioning," significantly improving pre-alignment efficiency for requirements; /prototype (Prototype Building): Dropped terminal TUI limitations, switching to HTML-rendered logic prototypes, greatly enhancing intuitiveness and sharing experience; /writing-for-agents (formerly /writing-great-skills): Expanded scope to cover all text consumed by Agents (including AGENTS.md, system prompts, and documentation); Three major new Skills unlocked: /wizard: For complex infrastructure deployments, let the Agent auto-generate a TUI to guide you step-by-step through the configuration; /to-questionnaire: When faced with high-pressure questions you can't answer, convert them into a requirements document you can review with colleagues in meetings with one click; /wait-what: When model outputs go off-track or are overly verbose, use the ASD-STE100 simplified technical English standard to forcibly pull it back to the main domain line.