Order from Lab 833636

Coding agent gyms

Reinforcement-learning environments built from private repositories: for each task the repository at a base commit, the issue a person wrote before the fix, the human fix, the tests that fail before and pass after, and a container that builds from scratch and runs those tests offline. Each task must be proven to run by someone other than the seller, come from code that was never public, and carry the right to execute the code and train on it.

How answering works

Any number of people fill one order. The lab receives one dataset, in its own columns.

Your agent reads the order
It looks at what you hold and says which columns you can fill, and what is missing.
Mora checks what you send
Every record: no copies, no personal data, not already public, not on Mora from someone else.
The lab pays per record it accepts
The lab named the price. You fill as much of the order as you can.

The columns the lab wants

You do not need every column. Mora says which ones each dataset fills.

Column Type What it must hold

task_id, repo_id id The task and the repository it belongs to

repo_snapshot git bundle or tarball The private repository at base_commit, with the history before it

base_commit sha The commit the agent starts from

problem_statement text The issue or bug report as a person wrote it, before the fix existed

issue_created_at, fix_merged_at timestamp Read from the tracker and the git host, not typed by the seller; the issue must predate the fix

gold_patch diff The human fix that closed the issue, source files only

test_patch diff The tests added or changed with the fix, kept apart from the fix

fail_to_pass list of test ids Tests that fail before the fix and pass after it

pass_to_pass list of test ids Tests that pass before and must keep passing after

environment Dockerfile and lockfile, or image digest A container that builds from scratch with pinned dependencies and runs with no network

test_command text One command that runs the tests and returns one exit code

validation_log log Before and after test runs, three times, made by Mora or another third party

rights text Ownership warranty, third-party licenses in the tree, exclusivity, how many labs already hold it

Column	Type	What it must hold
`task_id, repo_id`	id	The task and the repository it belongs to
`repo_snapshot`	git bundle or tarball	The private repository at base_commit, with the history before it
`base_commit`	sha	The commit the agent starts from
`problem_statement`	text	The issue or bug report as a person wrote it, before the fix existed
`issue_created_at, fix_merged_at`	timestamp	Read from the tracker and the git host, not typed by the seller; the issue must predate the fix
`gold_patch`	diff	The human fix that closed the issue, source files only
`test_patch`	diff	The tests added or changed with the fix, kept apart from the fix
`fail_to_pass`	list of test ids	Tests that fail before the fix and pass after it
`pass_to_pass`	list of test ids	Tests that pass before and must keep passing after
`environment`	Dockerfile and lockfile, or image digest	A container that builds from scratch with pinned dependencies and runs with no network
`test_command`	text	One command that runs the tests and returns one exit code
`validation_log`	log	Before and after test runs, three times, made by Mora or another third party
`rights`	text	Ownership warranty, third-party licenses in the tree, exclusivity, how many labs already hold it

How much, in the lab's words

5,000 validated tasks from at least 150 private repositories, at most 60 tasks per repository, at least six languages (Python, TypeScript, Go, Java, Rust, C#); plus a separate 500 tasks exclusive to us as a test set. Paid per validated task, refund per task that fails on our rerun.

repositories
tasks
environment
validation
history
rights