GHSA-3gp5-q4jw-3v94
HIGHGHSA-3gp5-q4jw-3v94 is a high-severity (CVSS 8.1) CWE-863 vulnerability in @budibase/server. O3 Security confirms whether GHSA-3gp5-q4jw-3v94 is actually reachable in your code before you act, and blocks exploitation at runtime until you patch.
Budibase: Basic app users can exfiltrate stored REST datasource auth by rewriting datasource base URL
Exploitation Status
Proof-of-concept exploit code exists
- CISA’s SSVC triage found public proof-of-concept exploit code for this CVE, though no confirmed active exploitation.
- A successful exploit gives an attacker total control of the affected component, not partial access.
Exploitation and automatability from CISA’s SSVC triage for GHSA-3gp5-q4jw-3v94.
EPSS Exploitation Probability
EPSS (Exploit Prediction Scoring System) is a daily probability model maintained by FIRST.org. It estimates the likelihood a CVE will be exploited in production environments within the next 30 days, derived from real-world threat intelligence signals.
How urgent is this, really
GHSA-3gp5-q4jw-3v94 plotted by exploitation likelihood (EPSS) against impact (CVSS). The shaded corner — EPSS 50%+ and CVSS 7.0+ — is where this CVE doesn't sit, though severity or exploitability alone can still warrant action.
Where this sits among everything scored
Of 370,894 CVEs with a current EPSS score, this one falls in the < 10% band (highlighted). Real counts from FIRST.org, not a sample — log-scaled since the landscape is heavily right-skewed.
Real-World Exposure
How broadly this vulnerability is actually deployed: weekly install volume shows current usage, and reverse-dependency count shows how many other packages break if it stays unpatched.
@budibase/servernpmDescription
Summary
Budibase stores external REST datasource credentials server-side and documents that database credentials are applied server-side and are not exposed in the UI. The REST datasource implementation redacts stored Basic/Bearer/OAuth2 auth secrets before returning datasource data to clients. However, the single-datasource GET and PUT routes are guarded by generic TABLE READ, not by Builder/Admin permission or datasource-specific ownership/resource checks.
The built-in Basic app user role maps to the WRITE permission set, which includes table read/write and query write. A Basic user can therefore read an existing REST datasource, receive redacted authConfigs values, submit an update that changes only config.url while keeping the redacted placeholders, and trigger an existing saved relative-path REST query. During update, mergeConfigs() restores the old stored secret when it sees the redaction placeholder. During query execution, Budibase prefixes the attacker-controlled datasource config.url to the relative query path and applies the resolved stored auth headers. The result is server-side disclosure of the builder-configured REST Authorization secret to an attacker-controlled listener.
Source evidence
packages/server/src/api/routes/datasource.ts: datasource list/create/delete routes are onbuilderRoutes, butGET /api/datasources/:datasourceIdandPUT /api/datasources/:datasourceIdare inauthorizedRoutesguarded only byPermissionType.TABLEandPermissionLevel.READ.packages/server/src/api/routes/datasource.ts: the:datasourceIdroutes do not attach datasource-specific resource authorization.packages/backend-core/src/security/roles.ts: built-in Basic user maps toBuiltinPermissionID.WRITE.packages/backend-core/src/security/permissions.ts:WRITEgrantsREAD/EXECUTElevels and includesQUERY WRITEandTABLE WRITE.packages/server/src/api/controllers/datasource.ts:datasourceController.updatereads the stored datasource, mergesctx.request.bodyinto it, writes the result back, and returns a redacted copy.packages/server/src/sdk/workspace/datasources/datasources.ts:removeSecrets()redacts REST Basic/Bearer/OAuth2 secrets toPASSWORD_REPLACEMENT.packages/server/src/sdk/workspace/datasources/datasources.ts:mergeConfigs()restores the old stored auth-secret field when the update body sends the redaction placeholder for the same auth config.packages/server/src/integrations/rest.ts: relative REST query paths are prefixed with datasourceconfig.url.packages/server/src/integrations/rest.ts: REST execution resolves the selected auth config and applies the resulting auth headers to the outbound request.packages/server/src/api/routes/query.ts: saved query executionPOST /api/v2/queries/:queryIdis guarded byQUERY WRITE, which the Basic role has through theWRITEpermission set.
Reproduction outline
No production systems were tested. This is source-backed and has a local static verifier plus a proof helper for an already-running authorized instance.
- Deploy a current Budibase instance.
- As a builder/admin, create and publish an app.
- As the builder/admin, create a REST datasource with:
config.urlset to a benign legitimate API base URL.- a stored REST auth config containing a sentinel secret, such as a Bearer token
BUDIBASE_REST_TOKEN_SENTINEL.
- As the builder/admin, create a saved REST query that uses a relative path and that auth config.
- Add a non-builder Basic app user.
- As the Basic user, confirm negative controls:
- Builder-only datasource list/create/preview routes are denied.
- The user is not a builder/admin.
- As the Basic user, call
GET /api/datasources/{datasourceId}. The response returns the datasource and redacted auth placeholders, not the raw secret. - As the Basic user, call
PUT /api/datasources/{datasourceId}with the same redacted datasource body but withconfig.urlchanged to an attacker-controlled HTTP listener. - As the Basic user, execute the saved query with
POST /api/v2/queries/{queryId}. - Expected vulnerable result: the attacker listener receives the server-side REST request with the preserved stored Authorization material, even though the Basic user never knew the raw secret and should not be able to administer datasource credentials.
Local source verifier:
python3 docker-proofs/s60/verify_budibase_basic_user_datasource_source_path.py
Expected success line:
SOURCE_PATH_VERIFIED budibase_basic_user_datasource_rest_secret_exfil
Observed May 1, 2026:
origin/masterwas8e6bf89acf1f602f3334592c4c8cd14e79f5362a.- Latest release was
3.37.2from Apr 30, 2026. - The source verifier passed and confirmed the route, role, redaction, merge, URL-prefixing, auth-header, and saved-query execution conditions.
Proof-assist helper:
python3 docker-proofs/s60/proof_budibase_basic_user_datasource_update_rest_secret_exfil.py \
--base-url http://127.0.0.1:10000 \
--app-id <published-app-id> \
--datasource-id <rest-datasource-id> \
--query-id <saved-relative-rest-query-id> \
--cookie '<basic-user-session-cookie>' \
--expected-secret BUDIBASE_REST_TOKEN_SENTINEL
The helper does not start, stop, or delete containers/resources. It targets an authorized already-running instance, rewrites only config.url, captures the outbound Authorization material, and restores the original datasource by default.
Impact
This breaks the intended application-user versus builder/admin boundary for external REST datasource credentials. A Basic app user should be able to use published app functionality, but should not be able to administer datasource connection settings or extract builder-configured REST auth secrets. In a realistic internal-tool deployment, REST datasource auth configs often contain bearer tokens, API keys, Basic credentials, OAuth client secrets, service account tokens, or integration credentials for ticketing, CRM, ERP, security, and operational systems.
An attacker with only Basic app-user access to an app that uses an authenticated REST datasource can redirect future query traffic to an attacker-controlled endpoint and collect the preserved server-side Authorization header. This is distinct from public REST datasource SSRF issues because the core impact is stored credential disclosure across the role boundary, and it works with an external attacker-controlled URL rather than depending on internal-network reachability.
Remediation ideas
- Move
GET/PUT /api/datasources/:datasourceIdbehind Builder/Admin datasource permissions, or add datasource-specific resource authorization. - Do not allow non-builder app users to update datasource
config,authConfigs, base URL, default headers, or plugin connection settings. - Split non-sensitive datasource metadata reads from credential-bearing/admin datasource reads.
- Treat redaction placeholders as valid only in trusted builder/admin update flows.
- Consider rotating REST datasource auth secrets for affected deployments after patching.
Duplicate/nearby public issue notes
Public triage found known Budibase REST datasource SSRF and protected-endpoint auth-bypass CVEs, but no obvious public duplicate for this specific Basic app-user PUT /api/datasources/:id role-boundary issue combined with preserved REST authConfigs secret exfiltration through a changed datasource base URL.
Affected Packages
| Ecosystem | Package | Vulnerable range | Fix |
|---|---|---|---|
| 📦npm | @budibase/server | all versions | 3.39.0 |
Detection & mitigation playbook
Open-source dependencyDetect
Scan your dependency tree (package-lock.json, pnpm-lock.yaml, requirements.txt, go.sum, etc.) for @budibase/server. O3's reachability analysis confirms whether the vulnerable code path is actually invoked in your application, so you act on real exposure instead of every transitive match.
Fix
Update @budibase/server to 3.39.0 or later, then make sure no transitive (indirect) dependency still pins the vulnerable range — O3 confirms GHSA-3gp5-q4jw-3v94 is resolved across your whole dependency graph.
Workarounds
If you can't upgrade right away: gate or disable the affected feature, validate untrusted input at the boundary, and avoid passing attacker-controlled data into the vulnerable path. O3's runtime protection blocks exploitation in production as an interim safeguard until the upgrade lands.
How O3 protects you
O3 pinpoints whether GHSA-3gp5-q4jw-3v94 is reachable in your code and exactly where to fix it, then blocks exploitation in production at runtime until the patched version is deployed.
Tailored to GHSA-3gp5-q4jw-3v94. Runtime protection reduces exposure until a permanent patch is applied and verified — it complements patching, it doesn't replace it.
Frequently Asked Questions
Is GHSA-3gp5-q4jw-3v94 in your dependencies?
O3 detects GHSA-3gp5-q4jw-3v94 across npm dependencies and uses function-level reachability to confirm whether the vulnerable code path is actually reachable — not just present. No false positives.