PYSEC-2026-2707
Vulnerability from pysec - Published: 2026-07-13 15:15 - Updated: 2026-07-13 16:05Global Knowledge Base Enumeration via knowledge-bases Meta-Collection
Affected Component
Retrieval collection access validation:
- backend/open_webui/routers/retrieval.py (lines 2330-2355, _validate_collection_access)
- backend/open_webui/routers/retrieval.py (query endpoints, e.g. POST /query/doc)
Affected Versions
Current main branch (commit 6fdd19bf1) and likely all versions with the knowledge base subsystem.
Description
The _validate_collection_access function uses an incomplete allowlist that only enforces ownership checks for collections matching user-memory-* and file-* patterns. All other collection names pass through unchecked — including the system-level knowledge-bases meta-collection, which stores the IDs, names, and descriptions of every knowledge base on the instance.
Any authenticated user can query this meta-collection directly via the retrieval query endpoints to obtain a global index of all knowledge bases across all users.
# retrieval.py:2330-2355 — incomplete collection allowlist
def _validate_collection_access(user, collection_name, ...):
if collection_name.startswith('user-memory-'):
# Check user-memory ownership
...
elif collection_name.startswith('file-'):
# Check file access
...
# Everything else (including "knowledge-bases") passes through unchecked
This finding is the enabler for the KB destruction (process/web), KB content injection (process/file), and RAG vector search access bypass findings — all of which require knowing a target KB's UUID. Without this enumeration, UUIDs are random and practically unguessable; with it, UUIDs across the entire instance are trivially obtained.
CVSS 3.1 Breakdown
| Metric | Value | Rationale |
|---|---|---|
| Attack Vector | Network (N) | Exploited remotely via API call |
| Attack Complexity | Low (L) | Single API call |
| Privileges Required | Low (L) | Requires any authenticated user account |
| User Interaction | None (N) | No victim interaction required |
| Scope | Unchanged (U) | Impact within the knowledge base boundary |
| Confidentiality | Low (L) | Discloses KB metadata (IDs, names, descriptions) across all users |
| Integrity | None (N) | No direct data modification |
| Availability | None (N) | No denial of service |
Attack Scenario
- Attacker (any authenticated user) sends:
POST /api/v1/retrieval/query/doc { "collection_name": "knowledge-bases", "query": "confidential" } _validate_collection_accessdoes not recognize theknowledge-basesprefix and lets the request pass.- The vector search returns the most relevant documents from the meta-collection — knowledge base records including their UUIDs, names, and descriptions — across all users on the instance.
- Attacker varies the query to enumerate more KBs:
"project","internal","private", etc. - Attacker now has a full target list for subsequent attacks (destruction, poisoning, content extraction).
Impact
- Information disclosure: KB names and descriptions may reveal sensitive project names, internal initiatives, or user activities
- Enabler for other attacks: Unlocks the following findings by supplying the required target UUIDs:
- KB destruction/poisoning via
process/web - Cross-user content injection via
process/file - RAG vector search access bypass in
retrieval/utils.py - Transforms these from theoretical (requires UUID guessing) to trivially exploitable (UUIDs enumerable)
Preconditions
- Attacker must have a valid user account
| Name | purl | open-webui | pkg:pypi/open-webui |
|---|
{
"affected": [
{
"package": {
"ecosystem": "PyPI",
"name": "open-webui",
"purl": "pkg:pypi/open-webui"
},
"ranges": [
{
"events": [
{
"introduced": "0"
},
{
"fixed": "0.9.0"
}
],
"type": "ECOSYSTEM"
}
],
"versions": [
"0.1.124",
"0.1.125",
"0.2.0",
"0.2.1",
"0.2.2",
"0.2.3",
"0.2.4",
"0.2.5",
"0.3.0",
"0.3.1",
"0.3.10",
"0.3.12",
"0.3.13",
"0.3.14",
"0.3.15",
"0.3.16",
"0.3.17",
"0.3.17.dev2",
"0.3.17.dev3",
"0.3.17.dev4",
"0.3.17.dev5",
"0.3.18",
"0.3.19",
"0.3.2",
"0.3.20",
"0.3.21",
"0.3.22",
"0.3.23",
"0.3.24",
"0.3.25",
"0.3.26",
"0.3.27",
"0.3.27.dev1",
"0.3.27.dev2",
"0.3.27.dev3",
"0.3.28",
"0.3.29",
"0.3.3",
"0.3.30",
"0.3.30.dev1",
"0.3.30.dev2",
"0.3.31",
"0.3.31.dev1",
"0.3.32",
"0.3.33",
"0.3.33.dev1",
"0.3.34",
"0.3.35",
"0.3.4",
"0.3.5",
"0.3.6",
"0.3.7",
"0.3.8",
"0.3.9",
"0.4.0",
"0.4.0.dev1",
"0.4.0.dev2",
"0.4.1",
"0.4.2",
"0.4.3",
"0.4.4",
"0.4.5",
"0.4.6",
"0.4.6.dev1",
"0.4.7",
"0.4.8",
"0.5.0",
"0.5.0.dev1",
"0.5.0.dev2",
"0.5.1",
"0.5.10",
"0.5.11",
"0.5.12",
"0.5.13",
"0.5.14",
"0.5.15",
"0.5.16",
"0.5.17",
"0.5.18",
"0.5.19",
"0.5.2",
"0.5.20",
"0.5.3",
"0.5.3.dev1",
"0.5.4",
"0.5.5",
"0.5.6",
"0.5.7",
"0.5.8",
"0.5.9",
"0.6.0",
"0.6.1",
"0.6.10",
"0.6.11",
"0.6.12",
"0.6.13",
"0.6.14",
"0.6.15",
"0.6.16",
"0.6.18",
"0.6.19",
"0.6.2",
"0.6.20",
"0.6.21",
"0.6.22",
"0.6.23",
"0.6.24",
"0.6.25",
"0.6.26",
"0.6.26.dev1",
"0.6.27",
"0.6.28",
"0.6.29",
"0.6.3",
"0.6.30",
"0.6.31",
"0.6.32",
"0.6.33",
"0.6.34",
"0.6.35",
"0.6.36",
"0.6.37",
"0.6.38",
"0.6.39",
"0.6.4",
"0.6.40",
"0.6.41",
"0.6.42",
"0.6.43",
"0.6.5",
"0.6.6",
"0.6.6.dev1",
"0.6.7",
"0.6.8",
"0.6.9",
"0.7.0",
"0.7.1",
"0.7.2",
"0.8.0",
"0.8.1",
"0.8.10",
"0.8.11",
"0.8.12",
"0.8.2",
"0.8.3",
"0.8.4",
"0.8.5",
"0.8.6",
"0.8.7",
"0.8.8",
"0.8.9"
]
}
],
"aliases": [
"CVE-2026-44557",
"GHSA-6c2x-gcp3-gp73"
],
"details": "# Global Knowledge Base Enumeration via knowledge-bases Meta-Collection\n\n## Affected Component\n\nRetrieval collection access validation:\n- `backend/open_webui/routers/retrieval.py` (lines 2330-2355, `_validate_collection_access`)\n- `backend/open_webui/routers/retrieval.py` (query endpoints, e.g. `POST /query/doc`)\n\n## Affected Versions\n\nCurrent main branch (commit `6fdd19bf1`) and likely all versions with the knowledge base subsystem.\n\n## Description\n\nThe `_validate_collection_access` function uses an incomplete allowlist that only enforces ownership checks for collections matching `user-memory-*` and `file-*` patterns. All other collection names pass through unchecked \u2014 including the system-level `knowledge-bases` meta-collection, which stores the IDs, names, and descriptions of every knowledge base on the instance.\n\nAny authenticated user can query this meta-collection directly via the retrieval query endpoints to obtain a global index of all knowledge bases across all users.\n\n```python\n# retrieval.py:2330-2355 \u2014 incomplete collection allowlist\ndef _validate_collection_access(user, collection_name, ...):\n if collection_name.startswith(\u0027user-memory-\u0027):\n # Check user-memory ownership\n ...\n elif collection_name.startswith(\u0027file-\u0027):\n # Check file access\n ...\n # Everything else (including \"knowledge-bases\") passes through unchecked\n```\n\nThis finding is the enabler for the KB destruction (`process/web`), KB content injection (`process/file`), and RAG vector search access bypass findings \u2014 all of which require knowing a target KB\u0027s UUID. Without this enumeration, UUIDs are random and practically unguessable; with it, UUIDs across the entire instance are trivially obtained.\n\n## CVSS 3.1 Breakdown\n\n| Metric | Value | Rationale |\n|--------|-------|-----------|\n| Attack Vector | Network (N) | Exploited remotely via API call |\n| Attack Complexity | Low (L) | Single API call |\n| Privileges Required | Low (L) | Requires any authenticated user account |\n| User Interaction | None (N) | No victim interaction required |\n| Scope | Unchanged (U) | Impact within the knowledge base boundary |\n| Confidentiality | Low (L) | Discloses KB metadata (IDs, names, descriptions) across all users |\n| Integrity | None (N) | No direct data modification |\n| Availability | None (N) | No denial of service |\n\n## Attack Scenario\n\n1. Attacker (any authenticated user) sends:\n ```\n POST /api/v1/retrieval/query/doc\n {\n \"collection_name\": \"knowledge-bases\",\n \"query\": \"confidential\"\n }\n ```\n2. `_validate_collection_access` does not recognize the `knowledge-bases` prefix and lets the request pass.\n3. The vector search returns the most relevant documents from the meta-collection \u2014 knowledge base records including their UUIDs, names, and descriptions \u2014 across all users on the instance.\n4. Attacker varies the query to enumerate more KBs: `\"project\"`, `\"internal\"`, `\"private\"`, etc.\n5. Attacker now has a full target list for subsequent attacks (destruction, poisoning, content extraction).\n\n## Impact\n\n- **Information disclosure:** KB names and descriptions may reveal sensitive project names, internal initiatives, or user activities\n- **Enabler for other attacks:** Unlocks the following findings by supplying the required target UUIDs:\n - KB destruction/poisoning via `process/web`\n - Cross-user content injection via `process/file`\n - RAG vector search access bypass in `retrieval/utils.py`\n- Transforms these from theoretical (requires UUID guessing) to trivially exploitable (UUIDs enumerable)\n\n## Preconditions\n\n- Attacker must have a valid user account",
"id": "PYSEC-2026-2707",
"modified": "2026-07-13T16:05:11.228870Z",
"published": "2026-07-13T15:15:42.174376Z",
"references": [
{
"type": "WEB",
"url": "https://github.com/open-webui/open-webui/security/advisories/GHSA-6c2x-gcp3-gp73"
},
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2026-44557"
},
{
"type": "PACKAGE",
"url": "https://github.com/open-webui/open-webui"
},
{
"type": "PACKAGE",
"url": "https://pypi.org/project/open-webui"
},
{
"type": "ADVISORY",
"url": "https://github.com/advisories/GHSA-6c2x-gcp3-gp73"
}
],
"severity": [
{
"score": "CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:L/I:N/A:N",
"type": "CVSS_V3"
}
],
"summary": "Open WebUI vulnerable to Global Knowledge Base Enumeration via knowledge-bases Meta-Collection"
}
Sightings
| Author | Source | Type | Date | Other |
|---|
Nomenclature
- Seen: The vulnerability was mentioned, discussed, or observed by the user.
- Confirmed: The vulnerability has been validated from an analyst's perspective.
- Published Proof of Concept: A public proof of concept is available for this vulnerability.
- Exploited: The vulnerability was observed as exploited by the user who reported the sighting.
- Patched: The vulnerability was observed as successfully patched by the user who reported the sighting.
- Not exploited: The vulnerability was not observed as exploited by the user who reported the sighting.
- Not confirmed: The user expressed doubt about the validity of the vulnerability.
- Not patched: The vulnerability was not observed as successfully patched by the user who reported the sighting.