How it works

Lazarus is a Python package on the Claude Agent SDK with a Docker-backed sandbox. A resurrection runs through six organs.

Organ	Role
Scout	Reads a bare repo URL + its paper (web-enabled, but blind to your notes) and drafts the plan: capability, base image, and a falsifiable sanity check — so a revival starts from a link, not a hand-written goal.
Sandbox	Disposable container (CPU or GPU); every expensive success is snapshotted so a later failure never re-pays the build.
Commit-era pinner	Reconstructs the dependency universe as it was on the repo's last commit — the reasoning that beat the cu111/KeOps/cppyy tangle.
Repair loop	build → run → read traceback → patch → retry, bounded, isolated to the container.
Capability locator	Finds where "input → the famous output" happens and carves the minimal path to it.
Contract emitter	Module + CLI + pinned container + smoke test — CPU or GPU, verified callable on its own.

The integration contract

Every revival emits the same artifact — that uniformity is what makes a revived tool a composable brick. A contract (lazarus.yaml) declares a pinned base image, typed inputs/outputs, a containerised entrypoint using $INPUT / $OUTDIR, a smoke test (a metric + threshold you define), and — when the method reproduces a published number — a benchmark that emits a REPRODUCE.md certificate.

from lazarus.contract import Contract

c = Contract.from_yaml(open("examples/basset_predict_contract/lazarus.yaml").read())
print(c.name, c.base_image, c.smoke.metric, c.benchmark.measured)
# basset_predict lazarus/basset:site-ready mean AUROC... 0.8944

The Scout: reviving from a URL

The Scout is the one step allowed to see the outside world (it reads the public repo + paper via web tools) — but it is explicitly cut off from your local notes, memory, and project settings, so the downstream resurrection stays honest: it's solved from public information, exactly what a newcomer has. It emits a validated plan whose sanity check must be falsifiable (a metric + threshold, or an explicitly checkable qualitative assertion) — a revival that can't fail proves nothing.

Pointed at Basset (github.com/davek44/Basset) with no hints, the Scout picked the README-endorsed base image and wrote a benchmark-grade sanity check; the agent then reproduced the paper — and in doing so caught a silent soft-masking bug that a naive run would have shipped. The full gauntlet across all five repos is in the hard problems it solved.

Runs on a laptop, executes anywhere

Lazarus drives from your machine; where it executes is pluggable via one flag — a local container, a remote x86 box, a cloud VM, or a GPU rental. The agent's tools and the emitted predict.py both run against whatever --docker-host / DOCKER_HOST points at, so the whole chain is host-agnostic.