Build for a Friend: Receipt Lens
Build for a Friend: Receipt Lens Entered in the Best Use of Sentry Agent Tracing category: tooling that shows an agent's work, with traces. The problem A friend of mine runs AI agents to do real work: research, data pulls, summaries. The agents are useful. The problem is what happens after. The agent says "done," hands over an answer, and everything in between is a black box. He asked me the…
Receipt Lens is a tool that visualizes the execution trace of AI agents, showing every step they performed in order, how long each step took, and whether the receipt checks out cryptographically. The creator built this tool for a friend who runs AI agents for tasks like research and data analysis. The problem with these agents is that the steps they take are often a black box, making it difficult to verify if they completed the task accurately or if any steps were skipped or faked.
Receipt Lens addresses this by recomputing a cryptographic Merkle root from the step sequence and comparing it against the published value. If they don't match, it indicates that something changed after the fact, and the tool points out the specific check that caught the discrepancy. The tool also provides an execution timeline, highlighting slow steps, skipped steps, and any cryptographic verification errors.
It's built on an open IETF draft standard (AER-1), allowing anyone to implement and use the tool without needing licenses or permissions. The frontend is a single HTML page that allows users to paste a receipt or click demo buttons to test the tool. The backend is a single Python file using standard libraries only, making it easy to run.
The project was developed in just one weekend during the Hacktoberfest Weekend Challenge, demonstrating the power of open innovation.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
