INTEL_REPORT
arXiv — Cryptography & Security (cs.CR) · published 5/13/2026, 4:00:00 AM · TLP amber
Summary
Ingested excerpt (first ~500 chars of normalized text).
Sequential Behavioral Watermarking for LLM Agents arXiv:2605.11036v1 Announce Type: new Abstract: LLM-based agents act through sequences of executable decisions, but their trajectories provide little evidence of which agent or policy produced them, making provenance, ownership, and unauthorized reuse difficult to establish from observed behavior alone. This motivates watermarking signals embedded directly into agent behavior rather than only into generated text, since text w…
https://arxiv.org/abs/2605.11036
sha256:67cd6cbbc1ae2f3b5158d8db6b37c46668855763fa1926478b4267e3657c3805
What we pulled out
Deterministic extractor (IOC + allowlisted tokens + ATT&CK IDs present in DB).
Indicators
Linked with report → mentions → indicator. Values open the indicator workspace.
No indicators linked for this report.
Malware families
Allowlist token matches only.
Threat actors mentioned
Allowlist mentions — not a formal attribution verdict.
ATT&CK techniques
MITRE IDs referenced in text and present in local technique table.
CONTINUE INVESTIGATION
High-signal pivots without leaving the thread you started in search.
Browse the report corpus.