I Built TaskProof: Vet Bounties Before Wasting a Weekend

TaskProof is an open-source, browser-first tool designed to help freelancers and developers evaluate developer bounties before committing time and resources. It accomplishes this by converting a task brief into a checklist of required evidence, highlighting critical terms that may need clarification, and calculating the potential hourly return based on the advertised prize, platform fees, estimated build time, user-chosen chance of winning, and upfront costs.

For instance, if a bounty offers a $500 prize, a 7.5% platform fee, an estimated build time of 8 hours, and a user-assumed 15% chance of winning, the net prize if selected would be $462.50. However, after accounting for the 15% chance of winning, the probability-weighted figure would be $69.38, equating to approximately $8.67 per hour of speculative work. It's essential to note that the 15% winning assumption is not an AI prediction or measurement but a demonstration input for the hourly calculation.

TaskProof scans the task brief for specific phrases associated with upfront fees, wallet recovery requests, social account requirements, subjective judging, and time-sensitive deadlines. While these warnings serve as reminders rather than scam accusations, they can help users identify potential red flags. The tool generates a checklist that includes questions like the acceptance test, fund release process, reward division among winners, and whether the deadline follows local time zones.

To maintain privacy, TaskProof runs its AI component—FLAN-T5 Small, executed via Transformers.js and ONNX Runtime Web inside a browser's Web Worker—without sending user-pasted task briefs to external inference APIs. The deterministic rules and economics calculator function independently if the AI component fails. The tool is entirely self-contained, requiring no wallet connections, API keys, backend uploads, or sign-ups.

Although the project serves as an experiment, it could potentially benefit freelancers and small teams by providing an objective acceptance checklist, cash-out questions, and a reproducible work estimate. However, the author emphasizes that it is not a purchasable product and currently has no verified customers or earnings. The author has also maintained a free GitHub Actions Workflow Risk Check and offered a paid pilot service for reviewing GitHub Actions workflows.