AI Stack Alert Medium CVE-2026-126767 For AppSec, AI Eng, SOC / IR AI-assisted
CVE-2026-126767: OpsTree REMS through 1.0.0 prompt injection exposes hidden system prompt
OpsTree REMS through version 1.0.0 contains a prompt injection vulnerability. Authenticated users can submit crafted chat input to POST /api/v1/rca/chat to disclose the hidden system prompt. The disclosed internal orchestration instructions can aid crafting further prompt injection attacks.
Why it matters
Any authenticated user of an affected REMS deployment can read the hidden system prompt. Exposed instructions help an attacker design more effective prompt injection attacks against the AI assistant.
Do this
- Identify OpsTree REMS deployments at version 1.0.0 or earlier and check for a vendor fix or upgrade.
- Restrict access to the /api/v1/rca/chat endpoint to trusted users until a fix is applied.
- Review logs for POST requests to /api/v1/rca/chat that contain instructions to ignore rules or reveal system prompts.
- Move sensitive logic out of the system prompt and enforce access controls outside the model.
Exploit Clock
Exploitation unlikely (for now)
- EPSS 0.34%
Signals
- NoCISA KEV
- 5.3CVSS
- 0.34%EPSS
- NoPublic PoC
- 24/100Priority
- 60%Confidence
Evidence
“OpsTree REMS through 1.0.0 contains a prompt injection vulnerability that allows authenticated users to disclose the hidden system prompt”
Sources
Affected / related
OpsTree REMS
Published 2026-11-11 19:17 UTC · updated 2026-11-12 16:46 UTC · AI-assisted summary, checked against quoted sources · How we score