GHSA-G5XV-MHGM-V5F6
Vulnerability from github – Published: 2026-09-22 20:36 – Updated: 2026-09-22 20:36Summary
When OAuth tokens are saved, MCP Atlassian always writes a plaintext fallback copy under ~/.mcp-atlassian/oauth-<client_id>.json. The fallback file is created with the process default umask rather than restrictive permissions. In this environment the file was created as mode 0664, exposing access and refresh tokens to same-group local users and any process that can read the home directory.
Details
OAuthConfig._save_tokens() stores OAuth token data in keyring, but it also unconditionally maintains a plaintext file fallback for backwards compatibility in src/mcp_atlassian/utils/oauth.py:350-386. If keyring saving fails it also falls back to the same file path in src/mcp_atlassian/utils/oauth.py:387-391.
The fallback writer creates ~/.mcp-atlassian and then writes oauth-<client_id>.json with a normal open(token_path, "w") call in src/mcp_atlassian/utils/oauth.py:392-420. No mode=0o600, os.open(..., 0o600), chmod, or owner-only directory permission is applied. The file contains both access_token and refresh_token (src/mcp_atlassian/utils/oauth.py:360-367) and is later loaded from the same plaintext path in src/mcp_atlassian/utils/oauth.py:450-470.
The security policy warns that OAuth client credentials and secrets should not be exposed (SECURITY.md:39-44), but the current implementation creates a persistent plaintext token copy even when keyring succeeds.
PoC
The following safe local proof uses a temporary HOME and mocked keyring writes. It creates and deletes only temporary files.
uv run python - <<'PY'
import json, os, shutil, stat, tempfile
from pathlib import Path
from unittest.mock import patch
from mcp_atlassian.utils.oauth import OAuthConfig
home = tempfile.mkdtemp(prefix='mcp-atlassian-oauth-poc-')
old_home = os.environ.get('HOME')
os.environ['HOME'] = home
try:
cfg = OAuthConfig(client_id='poc-client', client_secret='client-secret', redirect_uri='http://localhost/callback', scope='offline_access', cloud_id='cloud-id')
cfg.access_token = 'poc-access-token'
cfg.refresh_token = 'poc-refresh-token'
cfg.expires_at = 2000000000
with patch('keyring.set_password', return_value=None):
cfg._save_tokens()
token_file = Path(home) / '.mcp-atlassian' / 'oauth-poc-client.json'
mode = stat.S_IMODE(token_file.stat().st_mode)
data = json.loads(token_file.read_text())
print(json.dumps({
'token_file_exists': token_file.exists(),
'token_file_mode_octal': oct(mode),
'contains_access_token': data.get('access_token') == 'poc-access-token',
'contains_refresh_token': data.get('refresh_token') == 'poc-refresh-token',
'token_file_path': str(token_file),
}, indent=2, sort_keys=True))
finally:
if old_home is not None:
os.environ['HOME'] = old_home
else:
os.environ.pop('HOME', None)
shutil.rmtree(home)
PY
Observed output from this environment:
{
"contains_access_token": true,
"contains_refresh_token": true,
"token_file_exists": true,
"token_file_mode_octal": "0o664",
"token_file_path": "/tmp/mcp-atlassian-oauth-poc-9m9wvktp/.mcp-atlassian/oauth-poc-client.json"
}
The proof confirms that a plaintext file containing both access and refresh tokens is created and is not owner-only.
Impact
A local user, container sidecar, compromised dependency, backup job, or other process with filesystem read access to the account's home directory can recover OAuth access and refresh tokens. Refresh tokens can allow continued Atlassian API access until revoked or expired, depending on the OAuth app and token policy. In shared hosts, Kubernetes volumes, developer workstations, and CI runners, this can lead to persistent Atlassian account compromise.
{
"affected": [
{
"package": {
"ecosystem": "PyPI",
"name": "mcp-atlassian"
},
"ranges": [
{
"events": [
{
"introduced": "0"
},
{
"fixed": "0.22.0"
}
],
"type": "ECOSYSTEM"
}
]
}
],
"aliases": [
"CVE-2026-77250"
],
"database_specific": {
"cwe_ids": [
"CWE-312"
],
"github_reviewed": true,
"github_reviewed_at": "2026-09-22T20:36:34Z",
"nvd_published_at": "2026-09-22T18:17:17Z",
"severity": "MODERATE"
},
"details": "### Summary\n\nWhen OAuth tokens are saved, MCP Atlassian always writes a plaintext fallback copy under `~/.mcp-atlassian/oauth-\u003cclient_id\u003e.json`. The fallback file is created with the process default umask rather than restrictive permissions. In this environment the file was created as mode `0664`, exposing access and refresh tokens to same-group local users and any process that can read the home directory.\n\n### Details\n\n`OAuthConfig._save_tokens()` stores OAuth token data in keyring, but it also unconditionally maintains a plaintext file fallback for backwards compatibility in `src/mcp_atlassian/utils/oauth.py:350-386`. If keyring saving fails it also falls back to the same file path in `src/mcp_atlassian/utils/oauth.py:387-391`.\n\nThe fallback writer creates `~/.mcp-atlassian` and then writes `oauth-\u003cclient_id\u003e.json` with a normal `open(token_path, \"w\")` call in `src/mcp_atlassian/utils/oauth.py:392-420`. No `mode=0o600`, `os.open(..., 0o600)`, `chmod`, or owner-only directory permission is applied. The file contains both `access_token` and `refresh_token` (`src/mcp_atlassian/utils/oauth.py:360-367`) and is later loaded from the same plaintext path in `src/mcp_atlassian/utils/oauth.py:450-470`.\n\nThe security policy warns that OAuth client credentials and secrets should not be exposed (`SECURITY.md:39-44`), but the current implementation creates a persistent plaintext token copy even when keyring succeeds.\n\n### PoC\n\nThe following safe local proof uses a temporary `HOME` and mocked keyring writes. It creates and deletes only temporary files.\n\n```bash\nuv run python - \u003c\u003c\u0027PY\u0027\nimport json, os, shutil, stat, tempfile\nfrom pathlib import Path\nfrom unittest.mock import patch\nfrom mcp_atlassian.utils.oauth import OAuthConfig\n\nhome = tempfile.mkdtemp(prefix=\u0027mcp-atlassian-oauth-poc-\u0027)\nold_home = os.environ.get(\u0027HOME\u0027)\nos.environ[\u0027HOME\u0027] = home\ntry:\n cfg = OAuthConfig(client_id=\u0027poc-client\u0027, client_secret=\u0027client-secret\u0027, redirect_uri=\u0027http://localhost/callback\u0027, scope=\u0027offline_access\u0027, cloud_id=\u0027cloud-id\u0027)\n cfg.access_token = \u0027poc-access-token\u0027\n cfg.refresh_token = \u0027poc-refresh-token\u0027\n cfg.expires_at = 2000000000\n with patch(\u0027keyring.set_password\u0027, return_value=None):\n cfg._save_tokens()\n token_file = Path(home) / \u0027.mcp-atlassian\u0027 / \u0027oauth-poc-client.json\u0027\n mode = stat.S_IMODE(token_file.stat().st_mode)\n data = json.loads(token_file.read_text())\n print(json.dumps({\n \u0027token_file_exists\u0027: token_file.exists(),\n \u0027token_file_mode_octal\u0027: oct(mode),\n \u0027contains_access_token\u0027: data.get(\u0027access_token\u0027) == \u0027poc-access-token\u0027,\n \u0027contains_refresh_token\u0027: data.get(\u0027refresh_token\u0027) == \u0027poc-refresh-token\u0027,\n \u0027token_file_path\u0027: str(token_file),\n }, indent=2, sort_keys=True))\nfinally:\n if old_home is not None:\n os.environ[\u0027HOME\u0027] = old_home\n else:\n os.environ.pop(\u0027HOME\u0027, None)\n shutil.rmtree(home)\nPY\n```\n\nObserved output from this environment:\n\n```json\n{\n \"contains_access_token\": true,\n \"contains_refresh_token\": true,\n \"token_file_exists\": true,\n \"token_file_mode_octal\": \"0o664\",\n \"token_file_path\": \"/tmp/mcp-atlassian-oauth-poc-9m9wvktp/.mcp-atlassian/oauth-poc-client.json\"\n}\n```\n\nThe proof confirms that a plaintext file containing both access and refresh tokens is created and is not owner-only.\n\n### Impact\n\nA local user, container sidecar, compromised dependency, backup job, or other process with filesystem read access to the account\u0027s home directory can recover OAuth access and refresh tokens. Refresh tokens can allow continued Atlassian API access until revoked or expired, depending on the OAuth app and token policy. In shared hosts, Kubernetes volumes, developer workstations, and CI runners, this can lead to persistent Atlassian account compromise.",
"id": "GHSA-g5xv-mhgm-v5f6",
"modified": "2026-09-22T20:36:34Z",
"published": "2026-09-22T20:36:34Z",
"references": [
{
"type": "WEB",
"url": "https://github.com/sooperset/mcp-atlassian/security/advisories/GHSA-g5xv-mhgm-v5f6"
},
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2026-77250"
},
{
"type": "WEB",
"url": "https://github.com/sooperset/mcp-atlassian/pull/1448"
},
{
"type": "WEB",
"url": "https://github.com/sooperset/mcp-atlassian/commit/b041733473f95119dd539542a43c280737a8e460"
},
{
"type": "PACKAGE",
"url": "https://github.com/sooperset/mcp-atlassian"
},
{
"type": "WEB",
"url": "https://github.com/sooperset/mcp-atlassian/releases/tag/v0.22.0"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:L/A:N",
"type": "CVSS_V3"
}
],
"summary": "MCP Atlassian: OAuth fallback token storage writes plaintext access and refresh tokens with group-readable permissions"
}
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.
The approach is described in our paper Mapping CVEs to MITRE ATT&CK Techniques: A Curated Gold-Set Classifier and the Limits of LLM-Assisted Label Expansion.
Browse all ATT&CK techniques and the vulnerabilities related to each.
Related by attack behaviour
Vulnerabilities whose description is nearest to this one in the vector space of the CIRCL/vulnerability-attack-technique-biencoder model. This is a similarity search over the bi-encoder space (plain cosine), not a classification, and it has no measured accuracy.