Borehole · Project · free tier
langchain-ai/langchain
Free reading: not for reliance
adf1e68 · 17,054 commits · 3,192 files
Read 2026-10-22

langchain-ai/langchain: technical due diligence, read from its repository

What Borehole read in langchain-ai/langchain's code and history at adf1e68: 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 & Operationsnot assessable
Fit with the Acquirernot assessable

What the repository says

  • Releases are cut from another branch than the one surveyed
  • .github/PULL_REQUEST_TEMPLATE.md:11: states a policy on AI-generated contributions

Architecture & Codebase — including AI provenance

strong
01.6

libs/partners/openrouter/langchain_openrouter/data/_profiles.py is 8,742 lines, 301× this codebase's median file

adequate

The median source file here is 29 lines. 40 files exceed 1,000 lines, the largest being 8,742. 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
libs/partners/openrouter/langchain_openrouter/data/_profiles.py:1 — 8,742 lines (301× median)
libs/core/langchain_core/runnables/base.py:1 — 6,730 lines (232× median)
libs/partners/openai/langchain_openai/chat_models/base.py:1 — 5,684 lines (196× median)
libs/partners/anthropic/langchain_anthropic/chat_models.py:1 — 3,384 lines (116× median)
libs/core/langchain_core/callbacks/manager.py:1 — 2,828 lines (97× median)

Engineering Process & SDLC Maturity

strong
02.16

Tests are retried until they pass

adequate

11 tests are individually marked to retry. A retry turns a flaky test green without saying why it failed, so a green build means the tests passed eventually. Ask how many retries a typical run needs, and which tests need them.

Evidence
libs/partners/openai/tests/integration_tests/chat_models/test_base.py:259 — test retry
libs/partners/openai/tests/integration_tests/chat_models/test_base.py:292 — test retry

Product & Engineering Maturity

strong
04.62

Releases are cut from another branch than the one surveyed

adequate

The newest pre-release, v0.1.17rc1, is not on the branch this survey read; the newest release on it is v0.1.16. The shipped product may differ from what this report describes. Survey the release branch too.

Evidence
tag v0.1.17rc1

Security & Reliability Posture

adequate
05.26

6 locked dependencies have published vulnerabilities, 31 high or critical in what ships

adequate

6 of 660 locked dependencies match 47 published advisories in OSV, 34 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
libs/partners/chroma/uv.lock — chromadb 1.5.9: GHSA-36p7-vc44-83pf (critical)
libs/partners/nomic/uv.lock — pillow 12.3.0: GHSA-332p-7hrc-92wf (critical), fixed in 13.0.0
libs/core/uv.lock — jupyter-server 2.20.0: GHSA-c3mw-737p-c7g2 (high), fixed in 2.21.0; a development dependency
05.65

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

adequate

The newest was published on 2026-06-12. A project that publishes advisories is telling its users about its fixes, and the count and pace say how much security work it carries. Whether 5 of them are fixed in the code you ship could not be told from this history: its version is only known to be at least 0.1.16. Answer lives in: the release you run, set against each fixed version.

Evidence
2026-06-12: GHSA-gr75-jv2w-4656 (moderate), fixed in 1.3.9: LangChain: Path traversal and sandbox escape in LangChain file-search middleware and loaders
2026-05-13: GHSA-3644-q5cj-c5c7 (high), fixed in 0.3.30: LangSmith SDK: Public prompt pull deserializes untrusted manifests without trust boundary warning
2026-05-05: GHSA-pjwx-r37v-7724 (high), fixed in 0.3.85: LangChain vulnerable to unsafe deserialization of attacker-controlled objects through overly broad `load()` allowlists
2026-04-16: GHSA-fv5p-p927-qmxr (moderate), fixed in 1.1.2: LangChain Text Splitters: HTMLHeaderTextSplitter.split_text_from_url SSRF Redirect Bypass
2026-04-16: GHSA-r7w7-9xr2-qq2r (low), fixed in 1.1.14: langchain-openai: Image token counting SSRF protection can be bypassed via DNS rebinding

Observability & Operations

not assessable

Not rated: only 5 of 16 checks could reach a conclusion here; the rest do not apply to this repository or could not read it.

Fit with the Acquirer

not assessable

Not rated: structurally not assessable from a repository.

Read more, or read your own

The full report on langchain-ai/langchain: 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.