IBM API Connect Unauthenticated SQL Injection (CVE-2026-9074)

Technical Analysis

The vulnerability resides in the password reset endpoint of IBM API Connect, allowing an unauthenticated attacker to inject malicious SQL statements that are executed against the backend database. Because no authentication is required, exploitation can be automated and scaled across internet-facing instances, potentially enabling extraction of user credentials, session tokens, API keys, and configuration data stored in the database. Given API Connect's role as an API gateway, successful exploitation could allow attackers to pivot into downstream services or manipulate API routing and access controls. Organizations that expose LLM tools, RAG pipelines, or agent orchestration endpoints through IBM API Connect are directly at risk, since compromised gateway credentials or database records could expose API keys and secrets used by AI agents to authenticate to backend model services, making this agent-relevant.

Affected Systems

IBM API Connect versions 10.0.8.0 through 10.0.8.9; IBM API Connect versions 12.1.0.0 through 12.1.0.3; specifically the password reset functionality/endpoint of the Developer Portal or Manager UI

Indicators of Compromise

• No public IOCs available at this time; monitor for anomalous POST requests to password reset endpoints containing SQL metacharacters (e.g., ' OR '1'='1, UNION SELECT, SLEEP()) and unexpected database error responses from API Connect instances

Remediation Steps

1. 1

   Apply vendor patch

   Upgrade to the IBM-provided fixed version of API Connect as soon as it is released; monitor IBM PSIRT advisories for patch availability and apply immediately in test then production environments.

2. 2

   Restrict access to password reset endpoint

   Implement WAF rules or reverse proxy filtering to block SQL injection patterns targeting the password reset functionality until a patch is applied.

3. 3

   Rotate credentials and secrets

   Rotate database credentials, API keys, and any secrets stored in or accessible via API Connect, especially those used by downstream AI agent or RAG pipeline integrations.

4. 4

   Enable database activity monitoring

   Deploy or tune database activity monitoring/logging to detect anomalous queries, especially against user/authentication tables.

5. 5

   Conduct incident response review

   Review logs for evidence of exploitation prior to patching, including unusual database queries, unexpected account resets, or data exfiltration indicators.