What is a coding bounty, and which ones are open right now?

A coding bounty is money posted for one specific piece of software work, paid to whoever delivers code that meets the stated terms. Some pay a single developer for fixing an issue, some pay for proving a security flaw, and some rank many competing entries on one published score. The table below lists the code, security and automation bounties open on make.fun, largest reward first.

Open bounties in these three categories on make.fun: 1, with rewards adding up to $250, as of November 8, 2026. Every reward on this board is stated in USDC, and every close time is in UTC.

Three shapes a coding bounty takes

Issue bounties. Someone attaches money to an open issue in a repository, usually on GitHub. A developer sends a pull request that resolves it and claims the money; the person who funded the issue checks the work and pays if it meets the requirements. On Opire, for example, the bounty creator reviews the pull request and arranges payment, and several developers may work on the same issue at once, so filing a claim is not the same as being paid.

Security bounties. A company or protocol publishes a scope and a reward for each severity level, then pays for valid vulnerability reports inside that scope. The job is finding and demonstrating a flaw rather than shipping a feature. HackerOne pays researchers by bank transfer, PayPal, or a USDC or Bitcoin wallet once they have verified their identity and filed an approved tax form, and only for reports the program accepts as valid.

Graded contests. One task, many entrants, a fixed deadline, and a metric announced in advance that decides the order. make.fun's code bounties work this way: your program reads an input and writes an output, the grader runs it against hidden tests written from the sponsor's brief, and the paid places go by rank on the result. No reviewer picks a favourite; the score sets the order.

Each shape rewards a different habit. An issue bounty favours someone who already knows the codebase and moves early. A security bounty favours depth on a single target. A graded contest favours an entry that holds up against tests you cannot see, which is why the terms matter more there than raw speed.

What to read before writing any code

1. The success test. For an issue bounty, the acceptance criteria written in the issue. For a contest, the exact metric and which direction wins. A make.fun code bounty names its kind: highest test score, first to pass every test, smallest program, fastest program, or cheapest to build. Each kind ranks entries differently.
2. Who decides. A maintainer, a security team, a panel of judges, or a grader that anyone can run again. The answer tells you how much of the outcome rests on one person's judgement.
3. The deadline and its time zone. make.fun deadlines are in UTC, and a posted bounty's entries window can be as short as 1 hour, so check the clock before you plan a weekend around one.
4. The hand-in format. A pull request, a written report, or a repository at one fixed commit. make.fun takes a public repository under an open-source licence together with the AI session transcripts that wrote it.
5. How the money moves. A card processor, a bank transfer, or a crypto wallet. make.fun pays in USDC on Solana to the winner's own wallet, which the winner claims after the result is final.

Why this board covers three categories

Code bounties fit the name of this site most closely, but make.fun's security and automation bounties are software work as well: a security bounty asks you to find a weakness under fixed rules, and an automation bounty asks for a program that finishes a job without a person stepping in. All three come down to code you write and hand in, so they share one table here. Trading, forecasting, growth and games bounties are left off; make.fun's own board lists those.

When the table is empty, nothing in these three categories is accepting entries on the day shown. Sponsors post new bounties on make.fun when they fund them, and this page is rebuilt from make.fun's public data with the date of each reading.

Working with an AI coding agent

Most bounty platforms leave your tools up to you and say little about them. make.fun goes the other way: an entry to any bounty with a reward must include the transcripts of the agent sessions that built it, from Claude Code, Codex or Cursor, and automatic checks confirm each transcript is real and matches the code. The companion page on how coding bounties are reviewed, paid and disputed explains what those checks look at, and how other platforms decide who gets paid.