Inspiration
Small competition teams need to understand how a captain's contribution scores translate into prize amounts, including rounding and the effect of changing a score. Transparent arithmetic can make that discussion more concrete without pretending the scores are objective.
What it does
A captain enters a CNY prize, member names and nonnegative integer contribution scores. ContriPact shows the score fraction, percentage and exact amount for each member. The largest-remainder method distributes every cent, with list order breaking ties. Changing an input marks the result stale and disables export. Recalculation compares amounts with the previous calculation. A JSON download includes the inputs, rule and both result snapshots.
How it was built
An independent React/TypeScript/Vite browser POC. The current release uses integer cents and BigInt arithmetic, with no backend, runtime AI or external API. Existing competition-specific work was retained while the active flow was narrowed to this deterministic slice. Only the product idea is reused from the earlier private product.
Engineering challenges and validation
The main implementation challenges were preserving every cent and preventing stale results from being exported. The solution uses BigInt quota arithmetic, deterministic largest-remainder rounding and immutable calculation snapshots. Automated tests cover normal shares, tails, zeros, invalid input, stale/tampered exports and 5,000 seeded conservation cases. Actual browser checks exercise modification, recalculation, before/after changes and a downloaded JSON. A fresh clone was installed, tested, built and run with Node 24.
Devpost Learn Skill Pack
Codex read the installed official 1-start through 6-ship pack and relevant templates. It resumed existing artifacts, then reopened scope, PRD and spec using the user's detailed supplied product answers before implementing the deterministic calculation/comparison/export slice. The root planning documents and current process record distinguish this revision from the previous AI prototype. The user delegated implementation; no new live interview, learner hands-on review or personal learning reflection is claimed. This project description was prepared with AI assistance.
Limitations
Scores remain human inputs: this tool does not prove objectivity, fairness, consent or legal agreement. Session data disappears on reload. It performs no payment or automatic AI scoring.
Log in or sign up for Devpost to join the conversation.