Summary
Flowise suffers from an authentication bypass in its password reset workflow, where the resetPassword method does not verify that a valid reset token was issued for the target account.
An attacker who knows a user's email address can call the reset-password API with a null or empty token and take over that account by setting a new password.
Affected context
This vulnerability appears in Flowise deployments exposing the /api/v1/account/reset-password endpoint, where the AccountService maintains password reset tokens as nullable or empty-string fields.
It becomes relevant when:
- Email addresses of valid users are guessable or can be enumerated
- The resetPassword logic accepts requests without verifying that a reset token was legitimately generated
- The application does not enforce additional verification factors in the password reset process
Exploit preconditions
- Attack surface: Publicly accessible HTTP API endpoint
/api/v1/account/reset-password - Required access: No prior authentication; only knowledge of a victim's email address is needed
- Trigger point: Invocation of the
resetPasswordmethod inAccountServicewith a null or empty reset token - Why exploitation is possible: The implementation assumes that the presence of an email address is sufficient and does not confirm that:
- A reset token was previously issued for that account, and
- The provided token matches a stored, non-null, non-empty, unexpired token
Why this happens
The core issue is an incomplete authentication and authorization check around the password reset operation, mapping to CWE-287: Improper Authentication.
Instead of treating the reset token as a mandatory proof that the user initiated the reset and controls the associated email channel, the code path allows null or empty tokens to proceed.
This creates a logic flaw where password reset becomes tied solely to knowledge of the email identifier, effectively bypassing the intended secondary authentication factor (the token).
Conceptually, the workflow fails to:
- Treat the reset token as a required credential
- Validate that such a credential was generated and bound to the account
- Reject resets when the credential is absent, invalid, or empty
Potential impact
- Full account takeover for any user whose email address can be guessed or discovered
- Unauthorized access to Flowise projects, flows, and configuration data
- Manipulation of LLM workflows, including injection of malicious logic or exfiltration paths
- Potential pivot into integrated systems, APIs, or data sources connected through Flowise
- Loss of confidentiality and integrity for data processed within compromised flows
General mitigation guidance
Effective mitigation centers on hardening the password reset workflow as an authenticated, token-driven process.
Common patterns include:
- Treating reset tokens as mandatory, high-entropy, single-use credentials bound to a specific account and expiry time
- Rejecting any password reset request where:
- No token exists for the account
- The provided token is null, empty, malformed, expired, or already consumed
- Decoupling the reset action from user identifiers alone; email/username must never be sufficient to authorize password changes
- Implementing rate limiting and monitoring for password reset attempts, especially for repeated requests against the same email
- Adopting secure-by-default identity workflows (e.g., signed reset links delivered over an out-of-band channel like email)
- Reducing privileges and visibility of administrative accounts to limit blast radius if compromise occurs
How teams detect issues like this
Detection efforts can focus on both code-level and runtime signals:
Static and code review signals
- Password reset methods that:
- Do not check for existence and validity of a reset token
- Accept null, empty, or default values for security-critical parameters
- Authorize sensitive state changes (like password updates) based solely on user identifiers
- Absence of checks for token expiry, revocation, and one-time use
- Lack of audit logging around password reset operations
Runtime and monitoring signals
- Logs showing frequent reset requests for many different accounts originating from the same IP or client
- Password resets executed without corresponding "reset token issued" events
- Unusual spikes in successful password resets, especially for privileged or rarely used accounts
Security teams can use application security testing tools, linters, and custom rules in CI/CD pipelines to flag insecure authentication flows and missing token validation logic.

