Policy Pulse - Week of February 1, 2026 | Issue #1

Write-ups and Disclosures
1

disclose.io Policy Pulse

Week of February 1, 2026 | Issue #1

Your weekly briefing on cybersecurity policy affecting vulnerability disclosure and security research.

Top Story

NIST Opens Critical Comment Period on AI Agent Security

The Center for AI Standards and Innovation (CAISI) at NIST has published a Request for Information seeking stakeholder input on securing AI agent systems. The RFI, published in the Federal Register on January 8, 2026, specifically targets security risks unique to agentic AI—systems capable of taking autonomous actions that impact real-world systems and environments.

Unlike typical software vulnerability assessments, this RFI focuses on novel risks from machine learning models embedded within AI agents, including indirect prompt injection, data poisoning, and techniques used to manipulate model outputs. NIST is seeking “concrete examples, best practices, case studies, and actionable recommendations” from practitioners who have deployed and managed agentic systems.

This is a significant opportunity for the security research community to shape federal AI security guidance. CAISI (formerly the US AI Safety Institute) will use responses to develop guidelines for mitigating agentic AI security risks—an area where red teaming and vulnerability research expertise is directly applicable.

Why it matters for VDP: AI agents introduce entirely new attack surfaces that traditional vulnerability disclosure frameworks weren’t designed to address. Input on this RFI could help establish security researcher roles and protections for AI red teaming activities.

Comment deadline: March 9, 2026 — Submit via regulations.gov under docket no. NIST-2025-0035.

This Week in Policy

Federal Strategy & Regulation

  • CISA Retires Ten Emergency Directives, Issues New F5 Directive — On January 8, CISA announced the retirement of ten Emergency Directives issued between 2019-2024, the highest number retired at one time. These include historic directives for SolarWinds (ED 21-01), Microsoft Exchange (ED 21-02), and Pulse Connect Secure (ED 21-03). Simultaneously, CISA issued ED 26-01 requiring mitigation of vulnerabilities in F5 devices by March 1, 2026. (CISA)

  • CISA and Partners Issue AI Security Guidance for Critical Infrastructure — New joint guidance warns of operational technology risks from AI implementations and urges stronger governance and safeguards across critical infrastructure sectors. (TechRepublic)

CVE & Vulnerability Programs

  • Europe Launches GCVE: A Decentralized Alternative to CVE — On January 7, 2026, the Computer Incident Response Center Luxembourg (CIRCL) launched db.gcve.eu, the Global CVE Allocation System. GCVE offers decentralized vulnerability numbering through GCVE Numbering Authorities (GNAs), addressing concerns about the traditional CVE program’s single-source dependency. The system maintains backward compatibility—CVE-2023-40224 can be represented as GCVE-0-2023-40224. (CyberScoop)

  • CVE Foundation Eyes Operational Launch — Following last year’s funding scare, the CVE Foundation continues building toward operational capability as a U.S.-based nonprofit seeking diversified funding. The Foundation emerged after CISA’s 11th-hour contract extension kept the MITRE-operated CVE program running. (CVE Foundation)

AI & Emerging Tech Security

  • NIST Cyber AI Profile Comment Period Closed — The comment period for NIST’s preliminary draft Cybersecurity Framework Profile for AI (NIST IR 8596) closed January 30, 2026. Over 6,500 individuals joined the community of interest, and NIST held a workshop January 14 to discuss the profile alongside the forthcoming SP 800-53 Control Overlays for Securing AI Systems (COSAiS). Next step: initial public draft expected later in 2026. (NIST)

  • Cloud Security Alliance Releases Agentic AI Red Teaming Guide — The CSA’s guide provides a comprehensive framework for testing vulnerabilities unique to autonomous AI agents, addressing prompt injection, tool misuse, privilege escalation, and cascading failures. Experts note traditional red teaming methods are insufficient for these complex, multi-LLM environments. (CSA)

Legal & Researcher Protections

  • DOJ Good Faith Policy Remains in Effect, But Unchanged — The 2022 DOJ policy directing federal prosecutors not to charge CFAA violations for good-faith security research continues to stand but has seen no recent legislative reinforcement. Advocates continue calling for comprehensive CFAA reform, noting the policy can be rescinded by future administrations and doesn’t address civil liability or state laws. (DOJ)

International Developments

  • Pall Mall Process Drafting Industry Guidelines for 2026 — The international multi-stakeholder initiative addressing commercial cyber intrusion capabilities (spyware) is drafting Industry Guidelines based on its 2024-2025 consultations. The Process explicitly recognizes “the benefit that good faith security research, vulnerability disclosure, bug bounties for cyber defensive purposes and penetration testing can have on cyber security defences.” (UK Gov)

Upcoming Deadlines & Events

Date Event Action
2026-03-01 CISA ED 26-01 F5 Directive Federal agencies report implementation status
2026-03-09 NIST CAISI AI Agent Security RFI Submit comments via regulations.gov (NIST-2025-0035)
2026-06-06 FAR Council IoT Cyber Trust Mark Deadline for FAR amendment steps
2027-01-04 Consumer IoT Cyber Trust Mark Federal vendor labeling requirement takes effect

Worth Reading

Policy Pulse is a weekly bulletin from disclose.io. Keeping the security research community informed on policy that affects our work.

Have a tip or want to contribute? Reply to this email or reach out on Twitter/X.