TS Evidence Graph: Make Every SKILL Instruction 100% Enforced

Developer Samchon has introduced @ttsc/evidence, a tool designed to solve the problem of AI coding agents ignoring project instructions. Despite detailed documentation like AGENTS.md, studies show that frontier models frequently fail to follow engineering principles, often prioritizing test-passing over code quality. The @ttsc/evidence package addresses this by transforming these instructions into mandatory compiler rules. Developers must now include @evidence annotations in their code, citing specific rules from their skill files. If a function fails to satisfy these requirements, the build process will stop, forcing the AI to either adhere to the guidelines or refactor the code until it complies. By integrating these checks directly into the TypeScript compilation process, the tool ensures that engineering standards are enforced consistently, significantly improving code reliability and reducing the need for manual oversight in AI-assisted development workflows.