Borehole · Project · free tier
zenml-io/zenml
Free reading: not for reliance
f3824df · 9,047 commits · 3,719 files
Read 2026-10-22

zenml-io/zenml: technical due diligence, read from its repository

What Borehole read in zenml-io/zenml's code and history at f3824df: its licence, security advisories, releases, review practice, tests, dependencies and operations. Every finding cites the file and line it rests on. The full report also covers the team and asks the questions the code cannot answer.

DimensionRating
Architecture & Codebase — including AI provenancestrong
Engineering Process & SDLC Maturitystrong
Product & Engineering Maturitystrong
Security & Reliability Postureadequate
Observability & Operationsstrong
Fit with the Acquirernot assessable

What the repository says

  • README.md:240: names Linux Foundation
  • README.md:242: names CNCF

Architecture & Codebase — including AI provenance

strong
01.6

src/zenml/zen_stores/sql_zen_store.py is 17,109 lines, 234× this codebase's median file

adequate

The median source file here is 73 lines. 46 files exceed 1,000 lines, the largest being 17,109. Size is not a defect on its own — some problems genuinely live in one place — but these are where merge conflicts, review fatigue and single-owner knowledge concentrate, and they are the first thing that slows an inheriting engineer down.

Evidence
src/zenml/zen_stores/sql_zen_store.py:1 — 17,109 lines (234× median)
src/zenml/client.py:1 — 9,782 lines (134× median)
src/zenml/zen_stores/rest_zen_store.py:1 — 5,696 lines (78× median)
src/zenml/cli/utils.py:1 — 4,235 lines (58× median)
src/zenml/zen_stores/zen_store_interface.py:1 — 3,520 lines (48× median)

Engineering Process & SDLC Maturity

strong

No finding.

Product & Engineering Maturity

strong

No finding.

Security & Reliability Posture

adequate
05.3

155 dependencies float, and no lockfile is committed

adequate

Dependency versions are declared as ranges and no lockfile is committed, so a build today does not necessarily produce what a build last month produced. That matters twice over in a transaction: the thing being bought cannot be reproduced exactly, and a compromised upstream release arrives without anyone choosing to take it.

05.27

The build runs third-party code it does not pin

adequate

4 third-party CI actions are referenced by a tag or branch rather than a commit, so they run whatever that tag points to on the day; 7 steps pipe a download straight into a shell. Tags have been moved to malicious commits in real supply-chain attacks, and a piped script is whatever the server returns that minute. Pinning to a commit is the fix, and it is small.

Evidence
.github/workflows/update-templates-to-examples.yml:46 — action pinned by tag
.github/workflows/update-templates-to-examples.yml:120 — action pinned by tag
.github/workflows/ci-slow.yml:145 — download piped into a shell
05.17

119 of 288 migrations cannot be rolled back

adequate

119 schema migrations (41%) have no working way back: an empty `downgrade`, a `down` that does nothing, or a data migration with no reverse. A bad deploy that includes one is recovered from a backup, not reverted. Ask when a restore was last rehearsed.

05.21

4 shell commands built from strings

adequate

4 places run a command through a shell (`shell=True`, `os.system`, `exec` with a template) rather than passing arguments. Where any part comes from outside, it is command injection. Ask where each one's input comes from.

Evidence
src/zenml/cli/integration.py:237 — command run through a shell
src/zenml/integrations/spark/step_operators/spark_step_operator.py:256 — command run through a shell
src/zenml/utils/pipeline_docker_image_builder.py:530 — command run through a shell
05.22

1 deserialiser that can run code

adequate

1 places read data with a deserialiser that executes what it reads (pickle). That is safe on data the system wrote itself and remote code execution on anything else. Ask where each one's input comes from.

05.34

1 of 2 Kubernetes workloads declare no resource limits

adequate

1 workload manifests set no CPU or memory limits. One pod that leaks memory or spins can starve everything else on its node, and the scheduler cannot plan capacity for what it cannot see.

Evidence
helm/templates/worker-deployments.yaml — workload with no limits
05.65

6 advisories against this project in the last 24 months, 2 high or critical

adequate

The newest was published on 2026-08-07. A project that publishes advisories is telling its users about its fixes, and the count and pace say how much security work it carries. None of them affects the surveyed version. The record is incomplete: NVD: 2 records name zenml with nothing tying them to this repository (CVE-2024-28424, CVE-2024-4680).

Evidence
2026-08-07: CVE-2026-68772 (high): ZenML 0.94.6 contains a remote code execution vulnerability in the CloudpickleMaterializer component that allows attackers with write access to a shared artifac
2026-07-24: CVE-2026-11922 (moderate): A vulnerability in zenml-io/zenml versions 0.57.0 through 0.94.2 allows an attacker to bypass rate-limiting on the `POST /api/v1/login` and self password-change
2026-07-21: CVE-2026-11876 (moderate): In zenml-io/zenml version 0.94.2, the `GET /api/v1/stack-deployment/stack` endpoint (`get_deployed_stack`) lacks proper RBAC authorization checks, allowing any
2025-10-05: GHSA-q92x-2x5g-h365 (moderate), fixed in 0.84.2: ZenML is vulnerable to Path Traversal through its `PathMaterializer` class
2025-03-20: GHSA-6gmf-2369-c76c (high), fixed in 0.68.0: ZenML unauthenticated DoS via Multipart Boundry

Observability & Operations

strong
06.9

No error reporter is wired in

adequate

Nothing in the service initialises an error reporter (Sentry, Rollbar, Bugsnag, Honeybadger or an APM agent). An exception in production reaches a person only if someone is reading the logs, which usually means a customer reports it first.

Evidence
no error reporter initialised

Fit with the Acquirer

not assessable

Not rated: structurally not assessable from a repository.

Read more, or read your own

The full report on zenml-io/zenml: every dimension, the questions for the room, and the audit log.

Survey a repository: free on public repositories. Private code is read on your own machine by a collector whose source you can read first.