The problem CookieJar solves

A petty cash drawer, also known as a till or kitty 🐈 is a small cash fund utilized by traditional businesses to pay for minor expenses.

A Cookie Jar is an on-chain petty cash jar that allows whitelisted individuals in a high-trust environment to take "cookies" for any expenses they might have incurred.

Any organization or group (or even one person!) can fund a Cookie Jar as an experiment to see how it will be utilized by a whitelisted group, setting up an interesting new capital allocation mechanism that's only possible onchain.

This can be used by any organization—nimble high-trust organs, sub-DAOs, working groups, and departments—in a variety of interesting ways! In the base case scenario, small expenditures no longer have to go through an onerous process of requesting a reimbursement—you simply take an amount and leave a note as to why. By bypassing any complex proposal processes, everyone in a trusted environment is happier and more efficient!

A second-order effect is that tasks that might seem burdensome due to the proposal process not being worthwhile can now be incentivized more efficiently, with rewards being easily claimed as cookies! Cookie Jars can also be bootstrapped for other novel experimental purposes—we're excited to see how organs will use them.

The Cookie Jar is a simple mechanism that helps strengthen teams. We want to trust each other to do things and ensure that the incentives are worthwhile. And who doesn't love a yummy cookie?

Just don't be a greedy cookie monster, because nobody likes that.

Challenges I ran into

Key technical hurdles included managing scope creep and making critical feature and architectural decisions. For example, debates arose whether to include blacklists, and significantly, how to implement the registry of deployed contracts. This involved considering options like dedicated registry contracts, adding a mapping directly to the factory, or implementing indexing. This decision point also tied into the broader challenge of balancing thorough engineering against the need to get a product out to gather user feedback. Specifically regarding the registry, we considered whether building an indexing implementation via Ponder or The Graph was worth the additional time compared to adding a mapping to the factory of deployed jars. Both options would achieve the same functionality; however, the indexer would be the more sustainable approach. Although indexing is the long-term solution, we decided to launch this version with a registry in the factory contract first to test the product idea and gather feedback for future iterations.