Borehole · Project · free tier
tabbyml/tabby
Free reading: not for reliance
21b2904 · 3,695 commits · 2,099 files
Read 2026-10-23

tabbyml/tabby: technical due diligence, read from its repository

What Borehole read in tabbyml/tabby's code and history at 21b2904: 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 Maturityadequate
Product & Engineering Maturityadequate
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
  • ee/LICENSE: Part of the code is under a separate licence: ee/
  • .gitmodules:1: 1 submodule(s) this survey does not read: crates/llama-cpp-server/llama.cpp

Architecture & Codebase — including AI provenance

strong
01.6

ee/tabby-webserver/src/service/auth.rs is 2,067 lines, 29× this codebase's median file

adequate

The median source file here is 69 lines. 3 files exceed 1,000 lines, the largest being 2,067. 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
ee/tabby-webserver/src/service/auth.rs:1 — 2,067 lines (29× median)
ee/tabby-schema/src/schema/mod.rs:1 — 1,974 lines (28× median)
ee/tabby-webserver/src/service/page.rs:1 — 1,198 lines (17× median)
01.15

135 errors caught and discarded in product code

adequate

135 places catch an error and do nothing with it, across 45 files (0.8 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.

01.33

A move from JavaScript to TypeScript began 39 months ago and is 77% done

adequate

522 product files are in the new language and 159 are still in the old one, 39 months after the first new file appeared. The new language's share of the lines rose from 68% to 98%, and moved 0 percentage points in the last year. Two languages for one job doubles what an engineer has to know, and the unmoved half is usually what nobody wants to touch. Ask whether finishing it is planned, and what it costs.

Evidence
159 old, 522 new

Engineering Process & SDLC Maturity

adequate
02.3

Tests cover a thin slice of the codebase

adequate

116 test files against 1,385 source files (8.4%). This does not measure whether the tests are good, only whether enough of them exist for the codebase to be changed safely by someone who did not write it — which is exactly the situation after an acquisition.

Evidence
116 test files / 1,385 source files

Product & Engineering Maturity

adequate
04.3

118 direct dependencies in a single manifest

adequate

ee/tabby-ui/package.json declares 118 direct dependencies. Each one is a maintenance obligation, a supply-chain surface and a potential licence question that transfers with the asset. The count says nothing about whether any given dependency was a good choice; it says how many such choices a buyer inherits.

04.19

2% of bug fixes add or change a test

adequate

Of 407 fixes in the last 24 months, 8 touched a test. A fix without the test that would have caught the bug can quietly come back, and the suite never learns from what went wrong.

Evidence
8 of 407
04.62

Releases are cut from another branch than the one surveyed

adequate

The newest release, v4.3.1, is not on the branch this survey read; the newest release on it is v0.8.0. The shipped product may differ from what this report describes. Survey the release branch too.

Evidence
tag v4.3.1

Security & Reliability Posture

adequate
05.1

2 credential-shaped strings in the tree sit in tests, examples, documentation or local development settings

adequate

Strings shaped like credentials appear only where fakes are usual: test fixtures, examples, documentation, placeholders and local stacks. They are almost always fakes. Ask someone to confirm that none of them works anywhere, because a real key in a development file is still a real key. 1 other match set aside: expired tokens, licence tokens a vendor verifies with a public key, or placeholder text where a key would go. None grants access to anything.

Evidence
2 cited: JWT, where fakes are usual
05.2

None of the standard security guard rails are configured

adequate

There is no disclosure policy, no automated dependency updates, no named code owners and no pre-commit configuration. None of these is load-bearing on its own. Their collective absence says that security posture here is whatever the current team happens to remember to do, which is not a posture that survives the team changing.

Evidence
checked 5 kinds of guard rail across 2,099 tracked files; 0 present
05.25

1 environment file is committed

adequate

.env is tracked in git. Environment files are where credentials usually live, and committing one usually means nobody decided how secrets reach the running application. Whatever is in it is in the history for good. Check 05.1 says whether it holds anything credential-shaped.

Evidence
.env — committed environment file
05.27

The build runs third-party code it does not pin

adequate

22 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. 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/ast-grep-lint.yml:13 — action pinned by tag
.github/workflows/autofix-pnpm.yml:32 — action pinned by tag
05.13

2 outbound calls can wait forever

adequate

2 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
experimental/prompt-rewrite-eval/evaluator.py:25 — outbound call without a timeout
experimental/prompt-rewrite-eval/evaluator.py:175 — outbound call without a timeout
05.26

120 locked dependencies have published vulnerabilities, 13 high or critical in what ships

adequate

120 of 3,967 locked dependencies match 217 published advisories in OSV, 108 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. This lockfile pins development tools alongside what ships and does not say which is which, so some of these may never reach a user.

Evidence
pnpm-lock.yaml — next 13.5.3: GHSA-2464-fhr4-w9xw (critical), fixed in 15.5.28
Cargo.lock — lettre 0.11.7: GHSA-4pj9-g833-qx53 (critical), fixed in 0.11.22
pnpm-lock.yaml — koa 2.15.3: GHSA-593f-38f6-jp5m (critical), fixed in 2.15.4; a development dependency
05.48

15 places insert HTML from a variable without sanitising it

adequate

15 places set HTML from a value that is neither a literal nor passed through a sanitiser. Where the value can come from a user, that is stored cross-site scripting, and it is the first thing a buyer's security review looks for in a frontend.

05.61

Part of the code is under a separate licence: ee/

adequate

A directory named for a paid edition carries its own licence. That code is usually not open source even where the rest is; what the buyer may do with each part differs.

Evidence
ee/LICENSE — licence: unrecognised

Observability & Operations

not assessable
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 tabbyml/tabby: 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.