what is tic fr?

TIC, short for Taken In Consideration, is a reasoning method packaged as a skill that can be loaded into AI agents to make them think more carefully before producing an answer. The idea behind it is that language models tend to default to the easiest, most obvious answer and often miss edge cases, spatial contradictions, or the physical consequences of what they suggest. TIC pushes the model through five steps, breaking a task down into its smallest parts, tracing every part through possible scenarios, cross checking where those scenarios overlap, testing the result against real world plausibility, and finally identifying gaps, all before it produces any output. The repository contains an early proof of concept comparing outputs generated with and without TIC across three areas: image generation, creative writing, and coding or system design. For image generation it compares prompts and resulting images to see which are more spatially coherent and physically plausible. For writing it compares two short stories from the same starting prompt to judge narrative consistency. For coding it compares two designs for a restaurant reservation system to see how well each covers edge cases like concurrency and failure handling. The test prompts and written responses in this proof of concept are in Italian, so an English speaking user would need their own AI agent to translate them. Both comparisons in each test used the same coding agent environment and the same underlying model, changing only whether the TIC skill was active, and the author reports that TIC produced noticeably better spatial reasoning in image prompts, stronger cause and effect logic in writing, and broader edge case coverage in coding designs, while noting the sample size is small. To use it, a person loads the included SKILL.md file into their AI agent and triggers it with a short phrase such as usa TIC or think with TIC. The project's license is listed as still to be decided.