Read 2026-11-03
eosphoros-ai/db-gpt: technical due diligence, read from its repository
What Borehole read in eosphoros-ai/db-gpt's code and history at 0cf4ebf: 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.
| Dimension | Rating |
|---|---|
| Architecture & Codebase — including AI provenance | strong |
| Engineering Process & SDLC Maturity | adequate |
| Product & Engineering Maturity | adequate |
| Security & Reliability Posture | adequate |
| Observability & Operations | strong |
| Fit with the Acquirer | not assessable |
Architecture & Codebase — including AI provenance
strongCommit messages are templated rather than written
adequate43% of commit subjects are exact repeats of another subject, and subject length varies by only 21.3 characters. A log written by someone explaining their reasoning to a future reader does not look like this. The log is therefore weak evidence of intent, whatever the code is like.
| 104× 'WEB API independent' |
| 35× 'feat(editor): ChatExcel' |
| 27× 'feat(Agent): ChatAgent And AgentHub' |
packages/dbgpt-app/src/dbgpt_app/openapi/api_v1/agentic_data_api.py is 5,441 lines, 69× this codebase's median file
adequateThe median source file here is 78 lines. 37 files exceed 1,000 lines, the largest being 5,441. 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.
packages/dbgpt-app/src/dbgpt_app/openapi/api_v1/agentic_data_api.py:1 — 5,441 lines (69× median) |
web/pages/index.tsx:1 — 5,226 lines (67× median) |
web/new-components/dashboard/DashboardEditor.tsx:1 — 4,693 lines (60× median) |
web/new-components/chat/content/ManusRightPanel.tsx:1 — 3,529 lines (45× median) |
packages/dbgpt-app/src/dbgpt_app/openapi/api_v1/dashboard/schemas.py:1 — 2,376 lines (30× median) |
A move from JavaScript to TypeScript began 40 months ago and is 57% done
adequate525 product files are in the new language and 403 are still in the old one, 40 months after the first new file appeared. The new language's share of the lines rose from 57% to 94%, and moved 10 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.
| 403 old, 525 new |
Engineering Process & SDLC Maturity
adequate9% of recent commits change more than 2,000 lines
adequate38 of 432 commits in the last 24 months changed over 2,000 lines of code, not counting reformatting, renames, vendored or generated code. A change that size is rarely reviewed line by line. Ask how they were reviewed, and what made them that large.
| 38 commits over 2,000 lines |
Product & Engineering Maturity
adequate19% of bug fixes add or change a test
adequateOf 225 fixes in the last 24 months, 43 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.
| 43 of 225 |
2 jobs are each done by more than one library
adequateWithin a single package: UI component kits: antd, mui; dates: dayjs, moment. Each duplicate is bundle size, two ways of doing the same thing in one codebase, and two upgrade paths to maintain. It is usually the residue of a migration that stopped halfway.
web/package.json — several libraries per job |
API v1 and v2 have been served together for 31 months
adequatev2 first appeared in Mar 2024, and v1 is still in the code 31 months later. Two versions mean two surfaces to secure, document and support. Ask who still calls the old one, and whether retiring it is planned.
pilot/server/api_v1/api_v1.py — v1 since Jun 2023 |
dbgpt/serve/agent/app/endpoints.py — v2 since Mar 2024 |
Security & Reliability Posture
adequate2 Dockerfiles run the application as root
adequateNo 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.
docker/allinone/Dockerfile — runs as root |
docker/base/Dockerfile — runs as root |
The build runs third-party code it does not pin
adequate7 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; 6 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.
.github/workflows/doc-image-publish.yml:41 — action pinned by tag |
.github/workflows/doc-image-publish.yml:44 — action pinned by tag |
.github/workflows/code-checks.yml:20 — download piped into a shell |
29 outbound calls can wait forever
adequate29 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.
packages/dbgpt-app/src/dbgpt_app/knowledge/_cli/knowledge_client.py:89 — outbound call without a timeout |
packages/dbgpt-core/src/dbgpt/agent/expand/resources/search_tool.py:45 — outbound call without a timeout |
packages/dbgpt-core/src/dbgpt/model/proxy/llms/spark.py:190 — outbound call without a timeout |
2 SQL statements built from strings
adequate2 places in product code build SQL by formatting or joining strings rather than passing parameters. In each, what the code shows going in is a table or column name, a constant, or something the line does not show, not a request's value. Ask whether any value reaches these strings, rather than a parameter.
packages/dbgpt-ext/src/dbgpt_ext/datasource/rdbms/conn_doris.py:430 — SQL built from a string |
packages/dbgpt-serve/src/dbgpt_serve/evaluate/service/fetchdata/benchmark_data_manager.py:917 — SQL built from a string |
6 shell commands built from strings
adequate6 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.
packages/dbgpt-core/src/dbgpt/model/cli.py:367 — command run through a shell |
packages/dbgpt-core/src/dbgpt/util/configure/markdown.py:566 — command run through a shell |
packages/dbgpt-core/src/dbgpt/util/configure/markdown.py:568 — command run through a shell |
51 locked dependencies have published vulnerabilities, 57 high or critical in what ships
adequate51 of 1,619 locked dependencies match 141 published advisories in OSV, 80 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.
web/package-lock.json — dompurify 3.4.15: GHSA-cg2g-xh64-v4q7 (critical), fixed in 3.4.17 |
web/package-lock.json — form-data 4.0.0: GHSA-fjxv-7rqg-78g4 (critical), fixed in 4.0.4 |
web/package-lock.json — hull.js 1.0.6: GHSA-q849-wxrc-vqrp (critical), fixed in 1.0.10 |
53 places insert HTML from a variable without sanitising it
adequate53 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.
packages/dbgpt-app/src/dbgpt_app/static/old_web/_next/static/chunks/1353-1dacbd59a5cf5fb8.js:1 — HTML from a variable |
packages/dbgpt-app/src/dbgpt_app/static/old_web/_next/static/chunks/1353-1dacbd59a5cf5fb8.js:1 — HTML from a variable |
packages/dbgpt-app/src/dbgpt_app/static/old_web/_next/static/chunks/9924-42c72dae9efe5ccd.js:17 — HTML from a variable |
8 advisories against this project in the last 24 months, 8 high or critical
adequateThe newest was published on 2025-03-20. 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.
| 2025-03-20: GHSA-3248-f932-c76p (high): DB-GPT vulnerable to Cross-Site Request Forgery |
| 2025-03-20: GHSA-3xq5-x4fj-rff7 (critical): DB-GPT vulnerable to Arbitrary File Upload with Path Traversal |
| 2025-03-20: GHSA-6xgj-c5fx-5v57 (high): DB-GPT Uncontrolled Resource Consumption vulnerability |
| 2025-03-20: GHSA-7gj6-22m4-qfhx (critical): DB-GPT Arbitrary File Write vulnerability |
| 2025-03-20: GHSA-8pwp-phcg-h36g (high): DB-GPT Path Traversal vulnerability |
Observability & Operations
strongNo error reporter is wired in
adequateNothing 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.
| no error reporter initialised |
2 of 4 deployed images carry no version, or `latest`
adequate2 images in the deployment configuration are named without a version or as `latest`. A redeploy can change the software without a commit, a rollback has nothing to roll back to, and nobody can say from the repository what is running.
docker-compose.yml:5 — image without a version |
docker/compose_examples/observability/docker-compose.yml:86 — image without a version |
Fit with the Acquirer
not assessableNot rated: structurally not assessable from a repository.
Read more, or read your own
The full report on eosphoros-ai/db-gpt: 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.