Borehole · Project · free tier
startrail-org/leann
Free reading: not for reliance
b415df8 · 653 commits · 320 files
Read 2026-10-23

startrail-org/leann: technical due diligence, read from its repository

What Borehole read in startrail-org/leann's code and history at b415df8: 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 provenanceadequate
Engineering Process & SDLC Maturityadequate
Product & Engineering Maturityadequate
Security & Reliability Postureadequate
Observability & Operationsadequate
Fit with the Acquirernot assessable

What the repository says

  • .gitmodules:1: 6 submodule(s) this survey does not read: packages/leann-backend-diskann/third_party/DiskANN, packages/leann-backend-hnsw/third_party/faiss, packages/leann-backend-hnsw/third_party/msgpack-c

Architecture & Codebase — including AI provenance

adequate
01.2

The repository begins with most of itself already written

adequate

The first commit carries 278,432 lines — 56% of every line ever added. Whatever happened before that commit happened somewhere this repository cannot show you, so the early history is not available as evidence of how the system was designed.

Evidence
2025-06-30T09:05:05+00:00; +278,432 across 1231 files; subject: 'Initial commit'
01.6

packages/leann-core/src/leann/cli.py is 4,026 lines, 21× this codebase's median file

adequate

The median source file here is 184 lines. 2 files exceed 1,840 lines, the largest being 4,026. 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
packages/leann-core/src/leann/cli.py:1 — 4,026 lines (21× median)
packages/leann-core/src/leann/api.py:1 — 1,878 lines (10× median)
01.15

68 errors caught and discarded in product code

adequate

68 places catch an error and do nothing with it, across 20 files (1.1 per thousand lines). Each is a failure the system will not report. They surface later as problems nobody can explain, and they are cheap to fix now.

Evidence
apps/base_rag_example.py:365 — error caught and discarded
apps/document_rag.py:49 — error caught and discarded
apps/document_rag.py:63 — error caught and discarded

Engineering Process & SDLC Maturity

adequate

No finding.

Product & Engineering Maturity

adequate
04.20

Fixes are 52% of recent work, up from 33%

adequate

In the last two quarters 52% of commits were fixes, against 33% in the quarters before. Rising fix work usually means quality debt is coming due, and it takes time that new work would have had.

Evidence
57%, 48%, 40%, 27%, 38%, 25%

Security & Reliability Posture

adequate
05.6

1 Dockerfile runs the application as root

adequate

No USER instruction switches away from root, so the process runs with full privileges inside the container. It is a one-line fix and an ordinary expectation, which is why its absence is worth asking about: it usually says more about whether anyone has reviewed the deployment than about this container specifically.

Evidence
docker/Dockerfile — runs as root
05.13

6 outbound calls can wait forever

adequate

6 HTTP calls use a client with no timeout (Python `requests` without `timeout=`, or Go's default client). When a dependency slows down, each one holds a worker until the connection gives up, and enough of them take the service down with the dependency.

Evidence
packages/leann-core/src/leann/chat.py:484 — outbound call without a timeout
packages/leann-core/src/leann/web_search.py:29 — outbound call without a timeout
packages/leann-core/src/leann/web_search.py:60 — outbound call without a timeout
05.22

6 deserialisers that can run code

adequate

6 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.

Evidence
packages/leann-core/src/leann/api.py:242 — code-executing deserialiser
packages/leann-core/src/leann/api.py:854 — code-executing deserialiser
packages/leann-core/src/leann/api.py:924 — code-executing deserialiser
05.26

28 locked dependencies have published vulnerabilities, 47 high or critical in what ships

adequate

28 of 242 locked dependencies match 120 published advisories in OSV, 49 of them rated high or critical. A lockfile says which version is installed, not whether the flawed code is reachable from this product, so each needs a look. Most are fixed by an upgrade, and the advisory names the version.

Evidence
uv.lock — anyio 4.11.0: GHSA-82r6-8w77-94w6 (critical), fixed in 4.14.2
uv.lock — certifi 2025.8.3: GHSA-6q48-6c7w-9gw2 (critical), fixed in 2026.10.19
uv.lock — nltk 3.9.1: GHSA-7p94-766c-hgjp (critical), fixed in 3.9.3
05.9

2 vendored components with no licence

adequate

Third-party code is committed into this repository without any statement of the terms it arrived under. Unknown terms are harder to deal with than inconvenient ones: an awkward licence can be complied with, an absent one cannot. Each of these needs its origin established before the code changes hands.

Evidence
packages/leann-backend-diskann/third_party/embedding.pb.cc — 1 tracked files, no licence file
packages/leann-backend-diskann/third_party/embedding.proto — 1 tracked files, no licence file

Observability & Operations

adequate
06.1

No telemetry or error reporting is wired in anywhere

adequate

None of 278 readable files references an instrumentation, metrics or error-reporting library. If this system is in production, nobody learns that it broke from the system itself — they learn it from a customer. That is a reliability finding before it is an observability one.

Evidence
searched 278 tracked text files for 16 known instrumentation libraries; manifests for metrics, tracing and error-reporting families; locked dependencies for the packages that emit; 0 matched
06.2

No structured logging library is in use

adequate

Nothing references a structured logging library. Logs that are written to be read by a person are not searchable by a machine, which means incident response here is proportional to how well someone remembers the codebase rather than to what the tooling can find.

Evidence
searched 278 files for 10 structured logging libraries; 0 matched
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 startrail-org/leann: 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.