CVE-2026-133631: FlowIntel Insecure Note Variable Resolution, Over-Permissive Chatbot API Access, and Credential Exposure in Process Arguments
FlowIntel contains multiple security flaws in its note variable resolution, chatbot API tooling, and AI report generation.
- Note variable references (e.g. cross-case or cross-task lookups) were resolved using the viewing user's privileges rather than the destination case's audience. An authenticated user with access to a shared or imported note could embed references to private cases, causing sensitive case data to be pulled into AI report requests, exports, and MISP sends without the referenced case's audience having granted read access.
- The chatbot API tool permitted the language model to invoke any FlowIntel API endpoint, including administrative routes and state-changing GET operations, without an allow-list or path validation. An attacker who can influence the model's instructions (e.g. via untrusted case content) could induce data modification or access restricted endpoints.
- AI report generation passed the Ollama bearer token and full case content as command-line arguments to a child process, exposing them in the system process table to any local user.
Affected versions: FlowIntel prior to 4.1.0.
Affected Software
Remediation
Recommended actions to resolve this vulnerability, in priority order.
- Upgrade
Upgrade to a fixed release to a version that resolves this vulnerability.
Fixed in 4.1.0
Event History
Frequently Asked Questions
Which users and deployments are exposed to the note-resolution issue?
Authenticated users who can access a shared or imported note are relevant to this issue. They could embed references to private cases or tasks, and the referenced data could then be included in AI report requests, exports, or MISP sends despite the destination case audience lacking read access.
What is required to exploit the chatbot API issue?
An attacker needs to be able to influence the language model's instructions, such as through untrusted case content. The chatbot tool could then invoke arbitrary FlowIntel API endpoints, including administrative endpoints and state-changing GET routes, because no endpoint allow-list or path validation was enforced.
Who can obtain the exposed Ollama token and case content?
Any local user able to view the system process table could potentially see the Ollama bearer token and full case content. These values were passed as command-line arguments when AI report generation launched a child process.
What versions should be remediated?
FlowIntel versions prior to 4.1.0 are affected. Upgrade to version 4.1.0 or later.