CVE-2026-61732
Decepticon: Role-boundary forgery via ChatML special-token literals in web crawl output composed into LLM context
Decepticon is an autonomous hacking agent for red teams. Versions prior to 1.1.17 wrap web crawl results — the output of agent reconnaissance against target services — into LLM messages without neutralizing ChatML special-token literals. Under the BYOK (Bring Your Own Key) deployment model, users configure their own LLM credentials to any OpenAI-compatible endpoint. Most open-source and self-deployed model providers (vLLM, SGLang, Ollama, LM Studio, text-generation-webui, etc.) do not filter special-token literals from user content in their default configurations. Those literals are parsed into structural role-boundary token IDs, meaning an attacker string planted in a target web page forges a new operator turn the model treats as authoritative, bypassing Decepticon's agent guardrails and resulting in arbitrary command execution inside the Kali Linux sandbox. Version 1.1.17 patches the issue.
- Affected
- BitterSecurity · Decepticon · decepticon-core · decepticon-sdk
- CVSS v3.1 GITHUB
- 10.0 CRITICAL
CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:C/C:H/I:H/A:H - EPSS
- 1.22% chance of exploitation in the next 30 days, 68th percentile
- Published
- 2026-09-24
Proof-of-concept exploits (2)
- advisories/GHSA-g5f9-3xfg-p9mf
- InertFluid/cve-2026-61732-lab0★ · 2026-09-29