{"skill":{"slug":"prompt-hardening","displayName":"Prompt Hardening","summary":"硬化 agent prompt、system prompt、SOUL.md、AGENTS.md、cron prompt 使 LLM 可靠遵循指令。触发词：agent 不听话、忽略规则、绕过约束、prompt 优化、指令合规、规则强化、prompt 硬化、LLM 不遵守、模型违规、creative circumve...","tags":{"latest":"1.0.0"},"stats":{"comments":0,"downloads":83,"installsAllTime":1,"installsCurrent":1,"stars":0,"versions":1},"createdAt":1775529732372,"updatedAt":1775531213831},"latestVersion":{"version":"1.0.0","createdAt":1775529732372,"changelog":"- Initial release of the prompt-hardening skill for systematically strengthening agent and system prompts to improve instruction compliance.\n- Provides 16 hardening patterns with usage recommendations, quick reference, and reliability ratings.\n- Includes CLI tool for prompt audits with a 16-point checklist.\n- Documents clear usage boundaries—focuses on prompt compliance, not code-level enforcement or code review tasks.\n- Delivers audit results, hardened prompts, and violation analyses as output artifacts.\n- Advisory skill only; changes are not applied automatically.","license":"MIT-0"},"metadata":{"os":null,"systems":null},"owner":{"handle":"lanyasheng","userId":"s179bk63dv96y90hw97xaay4fx83t7yr","displayName":"_silhouette","image":"https://avatars.githubusercontent.com/u/22570800?v=4"},"moderation":null}