Adversarial security review for AI skills, repos, MCP servers, and packages before you install them. A read-only scanner plus a five-persona reasoning pass, ending in a GO / GO WITH MITIGATIONS / NO-GO verdict.
Governance, skills, and hooks stack for running Claude (or any capable LLM) as an autonomous senior engineer on any software project — frontend or backend, any language. MIT.
Gated Claude x Codex build loop for Claude Code: Codex plans and implements from one persistent thread, Claude approves the plan and reviews the diff. Zero deps, 20-test smoke suite.
Scientific agent skills for Claude Code and Codex that turn research goals into auditable hypothesis generation, review, ranking, evolution, and synthesis.
繁體中文(台灣)去 AI 味 skill:60 種 AI 寫作特徵、157 組兩岸用語在地化、事實保留檢查(--source)、30 案例 eval 對照實測。Traditional Chinese (Taiwan) humanizer for Claude Code / Codex / Cursor, with a measurable no-fabrication check.