International legal context

Adherence to all treaties

Disarmament Treaties

Instrument	Status	Comment
1980 Convention on Certain Conventional Weapons (CCW)	Not Party
2001 Firearms Protocol	Not Party
Optional Protocol to the CAT	Not Party

International Humanitarian Law Treaties

Instrument	Status	Comment
1949 Geneva Conventions	State Party
Additional Protocol I of 1977	State Party
Additional Protocol II of 1977	State Party

International Criminal Law Treaties

Instrument	Status	Comment
1998 Rome Statute of the International Criminal Court	Not Party	Signed on 28 December 2000
Amendments to the Statute on the crime of aggression	Not Party
Amendment to the Statute on blinding laser weapons	Not Party

International Human Rights Treaties

Instrument	Status	Comment
1966 Covenant on Civil and Political Rights	State Party
Optional Protocol to the ICCPR	Not Party
1966 Covenant on Economic, Social and Cultural Rights	State Party
1948 Genocide Convention	State Party
1984 Convention against Torture	State Party
Convention on the Rights of the Child (CRC)	State Party
Optional Protocol to the Convention on the Rights of the Child on the involvement of children in armed conflict	State Party
2004 Arab Charter on Human Rights	State Party

Membership to International Organisations

Organisation	Status	Comment
United Nations	Member State
International Court of Justice	Did not accept compulsory jurisdiction
International Atomic Energy Agency (IAEA)	Member State	Safeguards agreement in place
League of Arab States	Member State

Domestic Legal Context

Imports, exports, and supplier base

According to the SIPRI Arms Transfers Database, the earliest transfers recorded span 1989–94. The Soviet Union supplied 2 Project-1241 missile boats, and 40 Strela-2 man-portable air-defence missiles and 20 P-15U anti-ship missiles (delivered in 1991). Russia supplied a further 100 Strela-2 missiles (1994); Ukraine delivered 4 Su-22 aircraft (1995); China supplied 3 Huangfen-class missile boats and 25 C-801 anti-ship missiles (1995); and North Korea supplied 45 Hwasong-6 surface-to-surface missiles (ordered in 1994, but not delivered until 2002).

Between 1998 and 2005: France supplied 15 AML armoured vehicles (delivered in 1999); Czechia supplied 132 T-55AM-2 tanks (2002); Russia supplied 39 T-72B tanks (2001); Belarus supplied 27 T-72B tanks (2000). Russia supplied 20 MiG-29SMT aircraft – 14 (2002) and 6 more (2005) – together with 100 R-27R/T air-to-air missiles (2002), 150 R-73 air-to-air missiles (2005), and 100 R-77 air-to-air missiles, 50 Kh-29 air-to-surface missiles and 60 Kh-31A1 missiles (all delivered 2005). Ukraine supplied 32 T-72 tanks and 100 BMP-2 infantry fighting vehicles (delivered 2003–04); Russia supplied a further 180 BMP-2s (2005).

Between 2005 and 2014: South Africa supplied 250 Qutaish-2 APCs (ordered 2008, delivered through 2014) and 50 Leopard armoured patrol vehicles (delivered 2010); Jordan supplied 10 Ratel-23 infantry fighting vehicles (2010) and 25 M-113 armoured personnel carriers (also 2010); Czechia supplied 15 OT-90 APCs (2010) and a further five in 2013; Belarus supplied 92 T-80 tanks (delivered 2012); and Saudi Arabia supplied 10 Shibl APCs (2010) – its first appearance as a supplier in the data reviewed. The United States supplied 32 M-113 APCs (2006), 18 Cougar MRAP vehicles (2008), and four UH-1H Huey-2 helicopters (2011). Russia supplied a further 100 BTR-80A infantry fighting vehicles (2010); Ukraine supplied 12 more Su-22s (2007) and 6 D-30 towed guns (2011).

From 2015, transfers recorded in the data supplied consist predominantly of armoured vehicles from Saudi Arabia or the UAE, alongside vehicles from unidentified suppliers and one aircraft transfer. The UAE supplied 25 M-ATV vehicles (in 2015), 25 Caiman APCs and 20 Cougar armoured patrol vehicles (2016), 50 Nimr-2 APCs (2018) and 25 MaxxPro APCs (both ordered and delivered 2018); Saudi Arabia separately supplied 25 M-ATVs (2015). The aircraft transfer consists of 3 AT-802U light attack aircraft, supplied by the UAE (ordered and delivered 2015) – the only fixed-wing aircraft recorded in the data supplied between the 2005–07 Su-22 deliveries and 2022. A further five Ratel-20 infantry fighting vehicles (2016) and 10 Kirpi APCs (2022) are listed with supplier given only as ‘unknown supplier(s)’. No deliveries are recorded in the SIPRI data reviewed for 2019–21 or after 2022.

SIPRI also records Yemen as a supplier, in each case before unification and in small quantities. South Yemen supplied Singapore with five Jet Provost armed trainer aircraft (1974) and four Strikemaster trainer/combat aircraft (1975); and supplied Ethiopia with 3 Project-205 missile boats (1986). No transfers with Yemen as supplier are recorded after 1986 in the data supplied.

Domestic legislation relevant to the Arms Trade Treaty

National framework

Law No. 40 of 1992 on Regulating Carrying Firearms and Ammunition and Their Trade is the principal identified Yemeni legislation governing civilian firearms and ammunition. There is no official English translation of the law identified in the sources reviewed. The English title used here follows the translation used by the Small Arms Survey. The law is supplemented by its Executive Regulation, issued by Republican Decree No. 1 of 1994.

Article 2 sets out the Law’s core institutional and definitional terms. It names the administering Ministry as the Ministry of Interior and Security, and the ‘Licensing Authority’ as the Minister or his delegate – the body responsible for civilian firearms licensing, carrying permits, and dealer oversight under the Law. The same article defines the weapons covered as personal firearms – rifles, automatic rifles, pistols, and hunting rifles – while providing that any weapon not falling within that definition is specific to the State’s military and security institutions.

Law No. 40 of 1992 primarily regulates the possession, carrying, and domestic trade of personal firearms and ammunition, consistent with this scope. Article 4 provides that the Law does not apply to weapons, ammunition, and storage facilities belonging to military and security institutions and used for defence and security functions, which remain governed by the laws, regulations, and rules applicable to those institutions instead; Article 50 similarly excludes government-issued weapons carried by members of the armed forces and security services in the course of their duties.

Article 24 provides that the State alone has the right to import firearms and ammunition of all types from abroad for defence and security purposes, through the competent official authorities, and that no natural or legal person may import or introduce firearms from abroad without State approval. Article 58 provides for the confiscation of weapons and ammunition smuggled into the country. For civilian firearms specifically, Article 25 of the 1994 Executive Regulation requires a valid licence for the importation or introduction of firearms and ammunition, subject to limited exemptions for specified individual hunting, sporting, and personal firearms.

The 1994 Executive Regulation also regulates the movement of firearms and ammunition within Yemen. Article 32 requires a permit from the Licensing Authority to move firearms or ammunition from one locality to another, including between governorates, specifying the quantity, origin, and destination, the consignor and consignee, and the route, timing, and security conditions of transport.

The legislation reviewed does not establish a comprehensive statutory framework. No provision was identified in Law No. 40 of 1992 or its 1994 Executive Regulation establishing a general licensing regime for the export, re-export or retransfer of military arms, or detailed statutory requirements concerning brokering or post-delivery verification; the available framework is directed principally at controlling the importation, possession, domestic movement, and trade of civilian firearms, while military and security holdings are left to separate laws and regulations governing those institutions.

For major conventional arms, no dedicated statute governing import, procurement, or end-use by the State has been identified. Article 24 establishes a general State monopoly over the importation of firearms and ammunition for defence and security purposes, but sets out no licensing criteria or oversight mechanism specific to major weapons systems, while Article 4 places military and security holdings outside the Law’s scope altogether. Article 4 suggests that military and security holdings are governed by other laws/regulations, but these were not identified.

Yemen has also reported measures relating to weapons accounting, storage and stockpile management in its implementation of the UN Programme of Action on Small Arms and Light Weapons. The training centre reported to UNIDIR is described under Transparency and reporting below. Yemen’s 2010 national report on implementation of the PoA also records the destruction of substantial quantities of explosives and ammunition and identified further needs concerning personnel training, security infrastructure, and control of the country’s coastline against illicit small-arms trafficking. These reports provide evidence of institutional practices and capacity-building efforts, but should not be treated as evidence of additional statutory transfer controls. 

A government enforcement campaign introduced in 2007 restricted weapons carrying in major cities, cancelled existing carry licences pending reissue under tighter conditions and targeted arms markets. The Small Arms Survey records a government announcement in October 2008 that 230 shops had been closed and 270 arms dealers arrested; it separately reports that as few as three markets remained open by late 2009. The 2007 campaign also included a short-lived buy-back of heavier weapons, including mortars, anti-tank shells, rocket-propelled grenades, mines and anti-aircraft guns.

A Yemeni government source cited by the Small Arms Survey put weapons seizures at 204,897 for 2008 alone – broadly consistent with, though not identical to, the 208,383 seizures the 2010 PoA report records for the longer window of 23 August 2008 to 30 November 2009. On civilian holdings generally, the Small Arms Survey’s 2007 estimate put small arms in Yemen at 6–17 million, averaging around 11 million – materially higher than the ‘no more than 8 million’ the Yemeni government itself reported in 2010 (see Transparency and Reporting, below).

No source examined assesses the current (post-2014/15) legal status or practical enforcement of Law No. 40 of 1992 and the 2007 campaign described above. The law is discussed as part of the published legal framework, while its practical application in IRG-controlled territory remains difficult to establish.

Transparency and reporting

No submissions to the UN Register of Conventional Arms (UNROCA) were identified in the material reviewed, and Yemen is not a party to the ATT. Yemen has submitted three national reports under the UN Programme of Action on Small Arms and Light Weapons (PoA): in 2003, 2006 and 2010.

UNIDIR’s analysis of the 2003 PoA reporting round leaves Yemen’s entries blank in its table of reported national legislation on manufacture and transit of SALW. Yemen’s 2010 PoA national report attributes the prevalence of weapons in Yemen to the wars that followed the September 1962 and October 1963 revolutions, the May 1990 unification, and the 1994 civil war; Yemen’s proximity to conflict in the Horn of Africa; its long coastline; its topography; and a tribal social structure in which weapons are considered part of masculinity. It identifies Law No. 40 of 1992 as the cornerstone of Yemen’s arms-control framework and notes a pending amending law before parliament.

Key Weaponry in the Armed Forces

Strategic position and defence posture

Yemen’s defence posture is defined by prolonged armed conflict, fragmented state authority, and a military architecture that remains heavily dependent on external support. The current conflict originated in 2014, when Ansar Allah (the Houthis) seized Sana’a, and escalated in March 2015 with the intervention of the Saudi-led coalition in support of the internationally recognized government. More than a decade later, the IRG, represented since April 2022 by the Presidential Leadership Council (PLC), is the sovereign in Yemen but it does not exercise territorial or military control over much of the country.

According to War WATCH’s 2026 assessment, three non-international armed conflicts (NIACs) were ongoing in or directly connected with Yemen: between the IRG and the Houthis; between the IRG and Al-Qaeda in the Arabian Peninsula (AQAP); and a separate but related NIAC between Saudi Arabia and the Houthis, revived in July 2026 after a multi-year hiatus. The IRG is therefore a party to two of the three current NIACs. A separate, short-lived conflict between the IRG and the Southern Transitional Council (STC), which began with the STC’s expansion into eastern Yemen in December 2025, ended following the Saudi-backed IRG counter-offensive in January 2026. The legal classification of the current conflicts and applicable IHL are addressed under Compliance with IHL and Human Rights by the Armed Forces, below.

The Houthis continue to control Sana’a and most of north-western Yemen, while the IRG exercises authority across much of the south and east through a heterogeneous collection of regular and affiliated military and security forces. AQAP retains an insurgent presence, particularly in southern and eastern Yemen, although its current operational capacity is substantially more limited than that of the Houthis. Territorial control nevertheless changed materially in September 2026, when a major Houthi offensive broke the relative stability of the principal front lines and advanced along the strategically important western coast. Ansar Allah gained control of Mokha, 75 kilometres from Bab al-Mandab, giving it a direct presence on the approaches to the strait, although fighting remained ongoing and the situation fluid.

A UN-brokered truce agreed in 2022 substantially reduced major combat between the Houthis and the IRG for more than four years. That did not, however, isolate Yemen from the wider regional confrontation. Houthi attacks against commercial shipping in the Red Sea and Israeli territory progressively internationalized the military environment. Israel first used force on Yemeni territory on 20 July 2024, subsequently conducting strikes against Houthi-controlled ports, Sana’a International Airport and power stations. In its assessment of that period, War WATCH classified the hostilities as an IAC between Yemen and Israel and as a separate NIAC between Israel and the Houthis. The United States and United Kingdom were likewise parties to a separate NIAC against the Houthis during that period; their air campaign intensified substantially in the second quarter of 2025. The increased frequency and intensity of the strikes produced an unprecedented peak in civilian casualties before direct US-Houthi hostilities ended following an Oman-mediated agreement on 6 May 2025. These conflicts are relevant to Yemen’s recent strategic environment but are not among the armed conflicts ongoing in September 2026.

The regional dimension remains central to Yemen’s current defence environment. Iran has provided the Houthis with military and technical support that has contributed substantially to the development of their missile, unmanned aerial vehicle, and maritime capabilities. The UN Panel of Experts on Yemen concluded in 2024 that, without foreign support, the Houthis lacked the capacity to develop and produce the range of complex weapons systems then available to them, including ballistic and cruise missiles, anti-ship missiles, surveillance and one-way attack UAVs and unmanned maritime systems. The Panel identified similarities between Houthi-operated materiel and systems produced or operated by Iran and other members of the ‘axis of resistance’, as well as evidence of external technical assistance, training, and weapons supply. Iranian support should not, however, be equated with Iranian control over Houthi operations: War WATCH similarly describes the Houthis as having close links with Iran but not being under its overall control.

This context acquired renewed importance in 2026. In July, Yemeni government forces, operating in coalition with Saudi Arabia, struck Sana’a International Airport to prevent an Iranian aircraft from landing, according to the Yemeni government. The Houthis subsequently attacked Saudi territory, including Abha airport, declared a naval blockade against Saudi-linked shipping in the Red Sea and attacked commercial vessels transiting Bab al-Mandab. These developments revived the previously dormant NIAC between Saudi Arabia and the Houthis; further Houthi missile and drone attacks against targets in southern Saudi Arabia followed in September. The escalation therefore reintroduced direct Saudi-Houthi hostilities as a distinct armed conflict within Yemen’s increasingly regionalised security environment, while large-scale fighting between the IRG and the Houthis was resuming inside Yemen.

For the IRG, these developments occurred during a fundamental reconfiguration of the anti-Houthi military camp. Government-held Yemen has long contained regular armed forces alongside military and security formations supported by different external patrons, principally Saudi Arabia and the United Arab Emirates, and operating with varying degrees of institutional and operational autonomy. That balance changed markedly at the end of 2025. The UAE-backed Southern Transitional Council (STC) seized large areas of Hadramawt and al-Mahrah in December, substantially weakening the IRG’s position in the south and east. A Saudi-supported counter-offensive in January 2026, involving Saudi airstrikes, Saudi-backed forces and local tribal forces, rapidly reversed those gains and restored government control over the contested territories. Although the STC disputed the fact of its dissolution, as Reuters reported on 10 January 2026, it ceased to operate as an effective fighting force following the January offensive and accordingly the short-lived IRG-STC NIAC ended in January 2026.

This did not, however, eliminate the fragmented military structures that had developed under competing Saudi and Emirati patronage. Rather, the January 2026 crisis initiated a broader reorganisation of government-side forces and altered the respective roles of the IRG’s principal external supporters. Following the STC’s defeat and the withdrawal of UAE forces, Saudi Arabia emerged as the dominant external sponsor of the government-side military architecture. According to the Sana’a Center, Riyadh moved to bring previously fragmented and, in some cases, formerly UAE-backed forces into a more coordinated structure, while strengthening its direct financial and operational support to the IRG.

A central instrument of this process is the Supreme Military Committee established by PLC Chairman Rashad al-Alimi to reorganize, equip, and unify military forces in IRG-controlled territory. Significantly, the Sana’a Center reports that the Committee formally reports to the Saudi-led coalition, illustrating the extent to which the restructuring combines Yemeni State institutions with direct Saudi oversight. Former Army chief of staff Taher al-Aqili was appointed Minister of Defence, while Riyadh moved to restructure forces that previously operated through different political/ external patronage networks.

The reorganization remains incomplete. The Sana’a Center’s April–June 2026 assessment describes an emerging structure characterized by substantial security hybridization:previously autonomous or externally supported formations are being brought into closer coordination with State institutions, but not all have been incorporated into the conventional military apparatus, and institutional and command relationships remain uneven. The restructuring has therefore increased coordination without creating a fully integrated national force. The constituent formations, their respective chains of command and their evolving institutional status are further examined under Composition of the Armed Forces, below.

The emerging arrangements should be understood against Yemen’s formal constitutional and institutional framework. Under the Constitution of Yemen, the State has exclusive authority to establish the armed forces, police and security forces, and no organisation, individual, group or political party may establish armed or paramilitary forces (Article 36). Article 38 establishes a National Defence Council, headed by the President, to address matters concerning the safeguarding of the Republic and its security, while Article 111 designates the President as Supreme Commander of the Armed Forces. Following President Abdrabbuh Mansur Hadi’s transfer of authority on 7 April 2022, these arrangements were modified by the Presidential Declaration establishing the Presidential Leadership Council (PLC). The Declaration transferred Hadi’s presidential powers to the PLC, assigned the Council responsibility for the country’s political, military and security administration during the transitional period, and required steps to integrate the armed forces under a unified national command structure. It separately vested the Chairman of the PLC with Supreme Command of the Armed Forces. The continuing fragmentation of the military establishment therefore remains at odds with this formal State-centred framework.

External dependence is also evident in equipment and operational capability. According to Janes, the Yemeni armed forces have ‘significantly benefited from the donation of military equipment by members of the Saudi-led coalition’, but this support has been only partially effective in addressing substantial capability gaps, leaving government forces dependent on foreign support, particularly Saudi air power. Janes assessed the PLC-aligned armed forces as ‘highly unlikely’ to possess advanced surface-to-surface missiles as of August 2026; its residual missile-related capabilities were principally anti-tank guided weapons and man-portable air-defence systems, alongside unguided rockets and a limited attack-UAV capability. The loss of substantial government weapons holdings during the collapse of State control in 2015, including the capture by the Houthis of the former Yemeni armed forces’ ballistic-missile arsenal, is addressed separately under Risk of Diversion and Re-export of Imported Weapons, below.

This operational and materiel dependence is reinforced by Yemen’s extremely constrained domestic resource base. Reliable measurement of current military expenditure is difficult. The World Bank’s SIPRI-derived military-expenditure series for Yemen ends in 2014, before the fragmentation of State institutions and public finances produced by the current conflict. The last available figure was approximately US$1.7 billion in 2014, equivalent to around 4 per cent of GDP. Contemporary defence resources therefore cannot be captured through a conventional national military-expenditure figure, particularly given the importance of external financing and formations supported outside ordinary State budgetary structures.

The wider fiscal context illustrates the constraint. The World Bank reported in June 2025 that real GDP per capita had fallen by 58 per cent since 2015 and that IRG revenues excluding grants had fallen to 2.5 per cent of GDP in 2024, in large part because of the continued suspension of oil exports. The economic situation remained severe in 2025, with the World Bank estimating in May 2026 a further contraction in real GDP and continuing severe fiscal constraints. These conditions substantially restrict the government’s independent capacity to finance salaries, equipment, maintenance, and military institutions.

External financing consequently forms an important part of the effective resource base sustaining government-side forces. Following the UAE’s withdrawal, Riyadh assumed financial responsibility for military and security personnel previously supported by Abu Dhabi. The Sana’a Center reported for January–March 2026 that Saudi Arabia announced a US$90 million package on 15 January to cover delayed public-sector salaries, including those of military and security personnel, followed by substantial additional development and budgetary assistance. By the second quarter, the scale of this dependency had become clearer: according to the Center’s April–June 2026 assessment, Saudi Arabia had budgeted nearly US$3 billion for 2026 to cover public-sector salaries in Yemen, including personnel previously paid by the UAE and affiliated with the STC.

There have also been efforts to restore more conventional fiscal governance. In early March 2026, the government enacted its first official State budget since 2019, but government finances remained heavily dependent on Saudi support. The coexistence of formal State expenditure, substantial direct foreign financing and military formations with differing institutional relationships makes it difficult to identify a consolidated defence budget or obtain a comprehensive public picture of military expenditure, procurement and force support.

These institutional and financial weaknesses operate within a wider conflict economy characterized by extensive illicit trade and smuggling networks. The Global Organized Crime Index 2025 gives Yemen an arms-trafficking score of 9.5/10, among the highest globally, and describes its illicit arms market as extensive and transnational, sustained by both maritime and overland smuggling routes and closely linked to the country’s prolonged conflict and fragmented governance. Weapons circulate domestically and across borders, including through routes in the Arabian Sea and Gulf of Aden towards the Horn of Africa, while porous borders, weak enforcement and corruption facilitate trafficking. The Index also identifies substantial illicit markets in oil and other commodities and describes armed groups and other State and non-State actors as exploiting Yemen’s fragmented governance and war economy.

Maritime insecurity provides a further indication of the weakness of State control over Yemen’s surrounding waters. During May and June 2026, piracy re-emerged off the Yemeni coast, with the Sana’a Center recording a successful hijacking and a series of armed approaches to vessels in the Gulf of Aden and Arabian Sea. These incidents were distinct from Houthi maritime operations but further illustrate the range of armed and criminal threats affecting the maritime environment around Yemen.

Broader governance weaknesses further complicate defence-sector oversight. Transparency International’s 2025 Corruption Perceptions Index scores Yemen 13/100, ranking it 177th of 182 countries assessed. While the index does not measure defence-sector corruption specifically, this wider governance environment is relevant to the wider institutional environment in which procurement, stockpile management, and end-use controls operate. Combined with the scale of Saudi financial and operational support, a clear determination of the extent to which current PLC military expenditure is financed through the Yemeni State budget, external assistance, or other arrangements is not possible.

Composition of the Armed Forces

For the purposes of this profile, identifying the IRG’s armed forces requires distinguishing between the formal Yemen Armed Forces within the Ministry of Defence (MoD) chain of command and the broader body of military formations operating as, or affiliated with, State forces outside that conventional structure. Under the formal constitutional framework described above, supreme command rests with the Chairman of the Presidential Leadership Council (PLC), while the conventional armed forces principally comprise the Yemeni National Army (YNA) and its Military Regions. Former army chief of staff Taher al-Aqili was appointed Minister of Defence in early 2026.

The formal armed forces do not, however, encompass the entire government military architecture. More than a decade of conflict and competing Saudi and Emirati patronage produced a heterogeneous collection of regular, paramilitary and locally organized forces with differing command structures and institutional affiliations. Saudi-led restructuring in early 2026 has sought to reduce this fragmentation and bring formerly autonomous or externally supported formations into closer coordination with State institutions, but without incorporating them all into the conventional MoD structure. The resulting architecture therefore remains hybrid.

Accordingly, this profile distinguishes three broad categories according to their institutional position and command structure: (1) forces within the conventional MoD chain of command; (2) other military formations established by, subordinated to, or increasingly incorporated within IRG State structures but remaining outside the conventional MoD apparatus; and (3) IRG-affiliated military formations retaining significant autonomous command structures. These categories describe current institutional and command relationships rather than political alignment alone. Their boundaries remain fluid where the restructuring is incomplete or available sources do not establish formal incorporation or command authority. Political affiliation, salary payments, external sponsorship, and operational coordination are relevant indicators, but none alone determines a formation’s institutional status or the attribution of particular conduct.

Forces within the MoD chain of command. The first category comprises principally the YNA and its Military Regions, whose equipment is discussed below. The 2026 restructuring has involved changes to the leadership of the conventional army alongside efforts to improve coordination with other government-side forces. New commanders were appointed to the First and Second Military Regions, among other positions. The continuing hybrid nature of the emerging structure is illustrated by overlapping appointments: the Sana’a Center reported in July 2026 that the newly appointed commander of the First Military Region was simultaneously in command of a division of the Nation’s Shield Forces (see below). The boundaries between the conventional army and other State military formations are therefore not always institutionally clear-cut.

Other State military formations outside the conventional MoD structure. The second category comprises military formations established by, directly subordinated to, or increasingly integrated within the IRG State architecture, but which the available evidence does not establish as components of the conventional MoD chain.

The Nation’s Shield Forces are the clearest example. Formalized by PLC Chairman Rashad al-Alimi in 2023, they operate under his direct supervision rather than the Ministry of Defence. Saudi-trained and funded, Nation’s Shield forces participated in the January 2026 reversal of STC territorial gains and were subsequently tasked, alongside the Giants Brigades, with securing Aden. Their leadership changed during the subsequent restructuring. The Sana’a Center continues to describe them as a paramilitary force operating under al-Alimi’s personal supervision rather than the MoD.

The Giants Brigades, led by PLC member Abdelrahman al-Muharrami (‘Abu Zara’a’), are a more hybrid case. Historically organized and supported by the UAE, they developed outside the conventional Yemeni military establishment and formed part of the Joint Forces operating on the West Coast. Following the UAE’s withdrawal in January 2026, the Giants joined the Saudi camp, were entrusted with security responsibilities in Aden, and increasingly coordinated with the army and other pro-government forces ‘behaving like State actors’, while continuing to operate outside the formal apparatus. The institutional status of several formerly STC-affiliated military formations has likewise changed since January 2026. ACLED’s February 2026 revision of its Yemen actor classification treats the Hadrami Elite Forces, Shabwani Defence Forces, Hazm Brigades, and Saiqa Brigades as State military forces operating under the post-January 2026 PLC structure rather than as STC forces.

These developments should be distinguished from the restructuring of former STC internal-security formations. Following the government’s recapture of Aden, the former Security Belt Forces were renamed the National Security Forces and absorbed former Aden Security and Lahj Security forces. The Ministry of Interior (MoI) has sought ultimately to integrate these formations into its police-like Special Security Forces; it has also pursued the integration of personnel from the Hadrami Elite, Security Support, and Hadramawt Protection forces for internal-security purposes. ACLED accordingly codes the National Security Forces as a State police actor, rather than as part of the military forces of Yemen. Their evolution is relevant to understanding the wider restructuring, but their policing functions and weaponry are addressed separately under the section below, ‘Key Weaponry in Law Enforcement Agencies’.

The STC itself is not treated as part of the IRG’s armed forces. It became a belligerent against the IRG during the short-lived NIAC of December 2025–January 2026. Although the announcement of its dissolution was disputed by some STC leaders and political mobilization continued, its military position was fundamentally altered by the January offensive. The status of former STC military and security formations must therefore be assessed individually according to their subsequent institutional and command relationships rather than inferred from either the STC’s military defeat or the disputed dissolution announcement. As noted above, several such formations now operate as State-aligned military or police forces, while their integration remains part of an ongoing restructuring process.

IRG-affiliated forces retaining autonomous command structures. The principal example is the Joint Forces on the West Coast, an umbrella formation comprising several armed factions, including the National Resistance Forces (NRF), also known as the Guards of the Republic, led by PLC member Tareq Saleh. The Joint Forces developed outside the conventional MoD structure and retain a degree of autonomous command. Tareq Saleh’s subsequent appointment to the PLC further formalized his political relationship with the IRG. For the purposes of conflict classification, War WATCH identifies the relevant party to the conflict with the Houthis as ‘Yemen (IRG) and affiliated forces (including militias composing the Joint Forces on the Western Coast)’ and further states that, despite their degree of autonomy, the Joint Forces are State organs owing to their official affiliation with the IRG. War WATCH assesses that they form part of the IRG and are thus treated as State organs.

External support. Saudi Arabia has become the principal sponsor of the restructuring of government-side military and security forces following the UAE’s withdrawal. The Sana’a Center reports that Riyadh assumed sponsorship of forces previously funded by Abu Dhabi and placed the restructuring process under close Saudi oversight, including through the Supreme Military Committee, which reports directly to the Saudi-led coalition. Saudi involvement extends beyond financing to training, appointments, force reorganization, and operational support, giving Riyadh substantial influence over the emerging government-side military architecture. The nature of Saudi involvement in particular military operations is addressed separately where relevant below. Saudi and IRG forces have also participated in coordinated or joint military operations; the respective roles of Saudi and Yemeni forces are identified below where the available evidence permits.

Army

Per Janes, since the onset of the civil war in 2014, Yemen’s State military apparatus has fragmented along political, tribal, and local power lines. Forces aligned with the IRG operate under the designation Yemeni National Army (YNA), but the YNA constitutes only a subset of the broader pro-government armed-actor landscape and does not exercise unified command over all of it; combat power is augmented by a range of paramilitary and militia formations – including the Giants Brigades, Nation’s Shield Forces, and Hadrami Elite Forces described above – that are politically aligned but remain largely outside formal YNA command and control.

The YNA’s principal mission is internal security and counterinsurgency against Ansar Allah, with secondary counterterrorism operations against AQAP and Islamic State-affiliated elements; the pre-2014 external-defence and border-security mandate persists nominally but is secondary in practice. The country was divided before the war into seven military zones under a ‘task force’/axis system that nominally persists, though some axes remain partly or wholly under Ansar Allah control; YNA forces hold bases in the southeast, forward positions near Ansar Allah-held terrain, and positions along the northern border with Saudi Arabia.

Sustained combat losses have caused significant matériel attrition, and domestic procurement and regeneration capacity is limited; the YNA is accordingly dependent on external partners – principally donations of small arms, light weapons, ammunition and light tactical vehicles. Janes assesses the YNA is likely to remain structurally dependent on external support and unlikely to transition to a sustainable, centrally directed national force. Compulsory military service, reintroduced in 2007 after an unsuccessful all-volunteer experiment, and endemic low morale – driven chiefly by non-payment of salaries and lack of promotion – remain long-standing structural weaknesses that the civil conflict has likely reinforced along political and ethno-tribal lines.

Air Force

The Yemeni Air Force (YAF) retains only a residual ability to generate air power and is unable to defend national airspace or conduct sustained air operations beyond a limited tactical capacity. Before the civil war, force generation was already constrained by chronic underinvestment, poor aircraft serviceability, and limited pilot training and readiness; these structural weaknesses were compounded following the 2014–15 seizure of key installations by Ansar Allah and aligned factions, and by the subsequent Saudi-led air campaign of Operation Decisive Storm, which resulted in widespread destruction of the YAF’s fixed- and rotary-wing aircraft, surface-to-air missile systems, radar sites and supporting infrastructure. The YAF’s current capability is accordingly restricted to a small number of serviceable aircraft provided or sustained through international partner support, with limited operational availability and endurance.

The YAF’s role is defensive counter-air, but this cannot be fulfilled in practice: most air-defence systems and combat aircraft – including the MiG-29 and Su-22 strike variants – were captured in the 2014–15 takeover or destroyed in the 2015 coalition air campaign, leaving airspace control, air defence, and major air-support missions to allied Gulf air forces rather than Yemeni operational command. The YAF’s remaining assets operate only in a limited tactical supporting role – light strike and basic reconnaissance for ground forces, where feasible.

No reliable open-source evidence points to new fixed- or rotary-wing procurement; loss of maintenance infrastructure and logistics support are severe constraints.

Most remaining air elements operate from Al-Anad Air Base, near Aden, which has remained the YAF’s hub even as some central airbases have been recaptured. The service has long experienced discipline and morale problems – including strikes and protests over unserviceable aircraft and sub-standard spares – and is believed to field only a limited number of competent pilots and technical personnel, some trained in the United States or allied Gulf states.

The equipment inventory set out below is, per Janes, based on data representing the YAF’s pre-civil-war organization and equipment; Janes states explicitly that this ‘does not presently reflect the YAF’, but that ‘the basic framework may be retained as a future template for the force’.

Janes’ functional ‘control of the air’ assessment records no dedicated YAF capability in defensive counter-air, offensive counter-air, strategic attack, counter-maritime (anti-surface or anti-submarine warfare), electronic warfare, air-to-air refuelling, airlift or aeromedical evacuation. Counter-land (interdiction and close air support) is very likely the YAF’s only extant capability: YAF pilots have conducted air strikes against non-State armed groups across the country using UAE-supplied IOMAX AT-802 light attack/ISR aircraft based at Al-Anad.

The air force significantly increased its tempo of combat operations against Ansar Allah in 2026; there is limited verifiable open-source information on precisely which units and platforms conducted these strikes, but Janes thinks that the air force has used reconstituted F-5 aircraft. Most airstrikes against Ansar Allah since 2015 have been carried out by coalition allies, principally Saudi Arabia and the UAE, rather than by the YAF. The YAF relies on foreign State actors and the Army for collecting, processing, analysing, and disseminating information to support targeting and decision-making.

Air-defence systems

No current YAF air-defence systems were identified in the sources reviewed. The service’s surface-to-air missile systems and radar sites were, per Janes, destroyed during Operation Decisive Storm in 2015 (see above); man-portable air-defence systems are instead held by the Army.

Unmanned Aerial Vehicles (UAVs)

No dedicated UAV capability was identified for the YAF in the sources reviewed; as noted above, the service’s ISR needs are instead met by foreign State actors and the Army.

Navy        

The Navy has limited capacity to independently generate and sustain maritime security across Yemen’s territorial waters and EEZ; a small, heterogeneous fleet, degraded shore infrastructure, and limited sustainment and training capacity confine operational employment largely to coalition-enabled tasks – chiefly port security, near-shore patrols, and interdiction of illicit maritime transfers to Ansar Allah – under Saudi-led coalition rather than independent Yemeni command. Yemen’s position astride the southern Red Sea and Gulf of Aden nominally gives the Navy a strategic role protecting international shipping lanes to the Suez Canal, a role reinforced by earlier US/UK/French counter-piracy capacity-building but at present exercised largely by coalition forces.

Ansar Allah’s 2014–15 seizure of Al-Hudaydah, the Navy’s former principal base, together with the loss of other port facilities, vessels, and infrastructure, has progressively degraded the service, and procurement and sustainment remain highly dependent on external partners. Janes assesses that non-State actors’ asymmetric maritime-attack capabilities – including remotely operated and explosive-laden craft – are likely to remain a challenge, and that reliance on external security providers will continue.

The Navy sits under the Ministry of Defence, distinct from the Coastguard (Ministry of the Interior); its headquarters relocated from Hudaydah to Aden, where Navy and Coast Guard forces are primarily based, with smaller outposts along the rest of the coast. The Naval Academy (Hudaydah) and Command and Staff College (Sana’a) historically supplied officer training; both are now inaccessible to the government, and the current state of naval professional military education is unclear.

The equipment inventory set out below reflected assets held prior to the civil war. The current conflict has led to the loss of an unknown number of naval platforms to rebel forces, as well as inoperability of others for want of spares and maintenance. The figures below should accordingly be read as a legacy inventory, not a current order of battle.

Recent and Current Operations

As described, Yemen is affected by three ongoing NIACs, of which the IRG is a party to two – against Ansar Allah (the Houthis) and AQAP – while the third, between the Houthis and Saudi Arabia, resumed in July 2026. The short-lived IRG-Southern Transitional Council (STC) NIAC ended in January 2026.

The defining development of recent times has been the return of large-scale fighting between the IRG and the Houthis after four years in which the principal frontlines had remained comparatively stable. Fighting increased during August before developing into a major Houthi offensive in early September. Houthi forces moved southwards at speed along Yemen’s western coastline and in Taiz governorate, catching IRG forces unprepared, with pro-government forces divided and poorly led. ACLED assessed the fighting between 3 and 7 September as the most serious escalation since the April 2022 ceasefire and recorded at least 276 fatalities in the fighting over those five days alone. The World Health Organization recorded 194 deaths and 884 injuries from hostilities in the Maqbanah and Jabal Habashi districts of Taiz governorate and along the Hodeidah and Taiz coastlines between 24 August and 6 September.

The offensive subsequently produced major territorial changes. The Houthis seized the strategically located port city of Mokha on 10 September, approximately 80 kilometres north of Bab al-Mandeb, before taking Perim Island on 11 September, placing them in control of territory immediately adjacent to the strategically important strait. Houthi forces also deployed to the Greater and Lesser Hanish islands farther north in the southern Red Sea by 14 September. The advance substantially altered a western front that had remained comparatively static since 2022 and threatened both the Taiz-Mokha corridor and government positions controlling the approaches to Bab al-Mandeb.

The Houthi advance was assisted in part by an AI-enabled deception operation, in which government communication lines were hacked and an AI-generated recording of Tareq Saleh was played ordering his troops to flee. It also records the possible use of Claude software in support of weapons development, while noting Anthropic’s claim that the actors concerned had not successfully fielded an operational weapon as a result. These developments accompanied the extensive use of more conventional capabilities, including ballistic missiles, drones, artillery, and ground forces, during the renewed offensive.

Government and affiliated forces responded with counter-operations concentrated particularly in Taiz governorate and the high ground overlooking the western coastal corridor. Saudi Arabia also became increasingly involved in supporting the IRG: ACLED recorded at least 15 Saudi strikes against the Houthis between 3 and 7 September, principally in Taiz and in support of IRG forces, while War WATCH similarly reported that Saudi Arabia was conducting airstrikes in support of the IRG as government forces sought to recover lost territory. Government forces reported recapturing Mount Qarfan on 23 September, a position overlooking valleys between al-Shamayatayn and al-Waziiya, followed by Jabal Bazel and further operations across al-Waziiya, Hayfan, Gurab, and other parts of Taiz by 27 September. These battlefield claims were not all independently verifiable, but indicated that government forces had moved from retreat to counter-operations during the second half of September.

At the same time, the IRG substantially increased the use of its own air power. According to Al Jazeera, citing the Yemeni military, Yemeni government warplanes carried out at least 15 strikes on 5 Septemberagainst Houthi positions, gatherings and vehicles in Taiz and Hodeidah, providing air support to government forces resisting the Houthi offensive towards the Red Sea coast and Bab al-Mandeb. This followed government airstrikes on Houthi missile and drone launch sites and weapons depots in Marib and al-Jawf on 25 July. Al Jazeera characterized the September attacks as a significant increase in the government’s use of its own air power.

Following the loss of Mokha, government forces also declared a military ‘kill zone’ encompassing Mokha, Jabal al-Nar, and parts of the Taiz–Mokha corridor, warning civilians to remain away from Houthi positions, equipment and movements. The Yemeni Air Force subsequently began strikes within the designated area, including against Houthi military equipment in Dhubab. The targeting and civilian effects of these operations are considered under the section ‘Compliance with IHL and Human Rights by the Armed Forces’ below.

By 1 October, government forces had stabilised but not reversed the principal territorial losses caused by the Houthi offensive. They continued to hold defensive positions across the Taiz mountains, Lahj, the Kahboub Mountains overlooking Bab al-Mandeb and Marib, where more favourable defensive terrain, reinforcements and increased firepower had prevented a further Houthi breakthrough. Government forces were also reporting extensive air, drone, artillery and other strikes against Houthi positions and reinforcements.

The resurgence of the IRG–Houthi conflict has occurred alongside the revival of the separate Houthi–Saudi Arabia NIAC. In July, an Iranian aircraft not authorized by the IRG was prevented from landing at Sana’a International Airport, an episode War WATCH identifies as the immediate precursor to the renewed Saudi-Houthi conflict. On 13 July, the Yemeni government publicly stated that its forces had struck the airport runway to prevent the aircraft from landing, a claim also reported in Al Jazeera. Human Rights Watch, following analysis of satellite imagery, described Yemeni government forces ‘in coalition with Saudi Arabia’ as attacking Sana’a airport with at least five munitions, three of which struck its only functioning runway. Its corrected account identifies Yemeni forces operating in coalition with Saudi Arabia, but does not allocate individual munitions or aircraft to either State. It therefore supports IRG involvement and coalition cooperation, without establishing which State conducted each strike. 

Within hours, the Houthis attacked Abha International Airport in Saudi Arabia with ballistic missiles and drones and subsequently declared a naval blockade against Saudi Arabia. Houthi forces also attacked commercial shipping transiting Bab al-Mandeb, including a Tanzanian-flagged vessel on 11 August in which six people were killed in the attack and a subsequent strike against rescuers. The confrontation escalated further in early September. ACLED recorded dozens of Houthi drones and ballistic missiles launched on 8 September against energy infrastructure and other civilian and military targets in Abha, Jizan, Najran, and Khamis Mushait, injuring 73 people. The Houthis described the attacks as retaliation for an alleged Saudi strike on a prison in al-Jawf that killed seven people, including a child. These operations form part of the separate Saudi-Houthi NIAC and are included here as context, principally because renewed Saudi military engagement has become increasingly intertwined operationally with the IRG’s campaign against the Houthis.

The IRG’s separate NIAC with AQAP also continues, although at substantially lower intensity. ACLED found that AQAP activity during the first seven months of 2026 fell by approximately 70 per cent, from 63 recorded events during the corresponding period in 2025 to 19 in 2026. The organisation nevertheless retained an estimated fighting force of around 2,000, principally entrenched in the mountainous area between Abyan and Shabwa, with cells and operatives also reported in Aden, Marib, Hadramawt, and al-Mahra. AQAP continued to conduct lethal hit-and-run operations against government and affiliated forces, while pursuing cooperation and non-belligerence with the Houthis. War WATCH therefore continues to classify the situation as a NIAC notwithstanding the comparatively limited level of current hostilities.

The military situation changed again in early October. By 2 October, Saudi Arabia and Yemeni government forces were preparing a substantially larger counteroffensive to push Houthi forces away from the Red Sea coast and the approaches to Bab al-Mandeb. Al Jazeera reported that Yemeni government and affiliated forces had already substantially intensified their operations, while the Saudi-led coalition stated it had conducted 97 operations targeting Houthi reinforcements on the Tur al-Baha and Taiz fronts using aircraft, drones, artillery, missiles and snipers. On 5 October, Yemeni government forces reported taking Dhubab and strategic positions around the strait, while government sources also claimed the recapture of Mokha; the full extent of these territorial gains was not immediately independently verified. 

Al Jazeera reported that Saudi-backed government forces claimed to have captured several strategic positions near Bab al-Mandeb and that government troops, supported by heavy Saudi airstrikes, had reached the outskirts of Mokha; the Houthis disputed the reported territorial gains, and their full extent remained independently unverified. The same offensive subsequently gave rise to allegations concerning the conduct of government-side forces towards Houthi fighters, which are examined below. Associated Press 13 October reporting described renewed encirclement of Taiz after the Houthi capture of Turbah on 4 October, restricted supply routes, and shortages of fuel and medical supplies. It reported government forces fighting south-west of the city to break the encirclement.

The renewed hostilities have also caused large-scale civilian displacement, particularly along the western coast and in southern governorates affected by the Houthi advance and subsequent government counter-operations. By 28 September, the International Organization for Migration (IOM) had recorded 145,308 people displaced across eight governorates since the escalation began in June 2026, including 81,444 in Taiz and 34,860 in Lahj. Displacement continued to increase during the October counter-offensive: according to the IOM, as reported by Reuters, more than 200,000 people had been displaced across ten governorates by 7 October 2026. These figures reflect the broader humanitarian consequences of the escalation rather than displacement attributable exclusively to IRG operations.

By 13 October, the escalation continued. Reuters’ account of 11 October, updated on 12 October, reported repeated Houthi attacks on Riyadh’s King Khalid International Airport, including a 10 October attack whose toll rose to thirteen dead and over 300 injured and an 11 October attack killing one person and injuring eleven. These attacks concern the separate Saudi–Houthi conflict. AP reported further Saudi strikes on Kamaran, Qanawis, and al-Jawf on 13 October.

Compliance with IHL and Human Rights by the Armed Forces

Scope of the compliance assessment

As set out above, the IRG is currently a party to two NIACs, against Ansar Allah (the Houthis) and Al-Qaeda in the Arabian Peninsula (AQAP). Common Article 3 to the Geneva Conventions and customary international humanitarian law (IHL) apply to both conflicts. Yemen is also a State Party to Additional Protocol II, which applies to the IRG–Houthi and IRG–AQAP NIACs under the conditions identified by War WATCH. The separate Saudi-Houthi NIAC does not involve the IRG as a party. Saudi conduct is therefore distinguished from the IRG’s own conduct, while operations supporting the IRG remain relevant to the assessment of its external military partners.

The present section focuses on the conduct of the Yemen Armed Forces and the other IRG military formations identified above. Given the hybrid character of Yemen’s security architecture, institutional labels are not necessarily authoritative. Conduct by a security or paramilitary formation is included here where it was acting in a military capacity in connection with an armed conflict. Conduct arising principally from policing is addressed separately under ‘Police Compliance with the Law of Law Enforcement and Human Rights’ below, even where the personnel concerned belong to a formation that also performs military functions. A formation’s position outside the conventional MoD chain does not, by itself, exclude its status as a Yemeni State organ or attribution of its conduct to Yemen.

Saudi and Emirati operations supporting the IRG are addressed under the relevant thematic headings, with incidents attributed to the actors identified in the sources. Their conduct is distinguished from that of IRG forces; external support alone is not treated as establishing responsibility for another actor’s conduct. Saudi operations belonging solely to the separate Saudi-Houthi NIAC fall outside this compliance assessment and are mentioned elsewhere as strategic context.

The assessment gives greatest weight to recent conduct, particularly developments in 2025 and 2026 through October, but does not apply a strict temporal cut-off. Earlier findings are included selectively where they concern particularly serious violations, demonstrate a recurrent or institutional pattern, remain materially relevant because of unresolved accountability, or otherwise assist in assessing present compliance. Historical findings are identified as such and are not treated as evidence that the same conduct continues, in the absence of more recent information.

Evidence is uneven across actors, geographical areas, and periods. Independent documentation is substantially more extensive for some parties to Yemen’s conflicts than for IRG forces, while the fragmentation of government-side forces and restrictions on access can make identifying the responsible formation difficult. The termination in October 2021 of the mandate of the UN Group of Eminent International and Regional Experts (GEE) also removed the principal independent international investigative mechanism dedicated to documenting violations by all parties in Yemen. The analysis below therefore draws, where available, on mutually reinforcing findings of UN mechanisms, the UN Panel of Experts on Yemen, the National Commission to Investigate Alleged Violations of Human Rights (NCIAVHR), international and Yemeni human-rights organizations, and other independently verified reporting. Differences in the volume of documented violations between actors are not, by themselves, treated as evidence of correspondingly different levels of compliance.

UNMHA also supplied local reporting on ceasefire incidents and explosive-hazard casualties in Hudaydah while its mandate remained in force. Its ceasefire-support role was distinct from the GEE’s human-rights investigations and did not provide comprehensive attribution or judicial findings concerning violations by all parties.

Civilian harm and the conduct of hostilities

Yemeni government forces and government-aligned armed formations have been implicated in unlawful civilian deaths and injuries throughout the conflict, although the scale, nature, and attribution of incidents vary considerably. Historical investigations identified indiscriminate shelling by Yemeni armed forces and affiliated UAE-backed formations, alongside extensive civilian harm caused by Saudi-led coalition air operations conducted in support of the IRG. More recent national and independent monitoring documents incidents involving government forces, although far more investigated civilian harm has been attributed to the Houthis. For the purposes of assessing IRG compliance, the conduct of regular government forces must be distinguished from that of affiliated formations operating under separate command structures and external military partners.

Historical findings: ground operations by Yemeni armed forces and affiliated formations (2015–2021). The UN Group of Eminent International and Regional Experts on Yemen (GEE) investigated attacks by parties to the conflict in its successive 2018, 2019, 2020 and 2021 reports. Its findings provide evidence of indiscriminate attacks and failures to protect civilians, including during operations involving Yemeni government forces and UAE-backed armed formations.

In its 2019 detailed findings (A/HRC/42/CRP.1, paras. 545–552), the GEE investigated three shelling incidents during the 2018 offensive in al-Hudaydah involving Yemeni armed forces and affiliated UAE-backed formations. In one, a projectile struck a family home in al-Durayhimi, injuring six family members, including children. In another, mortar shells struck two civilian houses in al-Durayhimi, killing six civilians, including three children, and injuring three others. In a third incident, a shell landed near a school and mosque in al-Hudaydah city, killing three girls and injuring two men. The Group identified no apparent military objective in the immediate vicinity of the latter attack. The GEE found reasonable grounds to believe that Yemeni armed forces and affiliated UAE-backed groups had launched indiscriminate shelling attacks and had failed to take adequate precautions to protect civilians. It considered that the acts could engage individual criminal responsibility for war crimes. These findings constitute particularly significant evidence of unlawful attacks involving government-side ground operations, although they do not establish that every incident was perpetrated by regular Ministry of Defence forces rather than affiliated UAE-backed formations.

The GEE (A/HRC/42/CRP.1, paras. 553–557) separately investigated the 2 August 2018 mortar attack on the fish market and al-Thawrah hospital in al-Hawak district, al-Hudaydah, which killed at least 41 civilians and injured 111. It considered the use of mortar fire in this densely populated area indiscriminate. Photographs of munition remnants suggested the possible use of Rheinmetall Denel 120 mm mortar shells, principally available to coalition forces. However, both government-side and Houthi forces possessed 120 mm mortars, and ammunition could have been captured or otherwise acquired. The Group considered it probable that mortars fired near the hospital came from the airport, where government-side forces were based (para. 555), but did not conclusively establish the perpetrator.

Independent investigations provide further evidence of apparent indiscriminate shelling by government-affiliated forces. Human Rights Watch reported in August 2017 that government-affiliated forces appeared to have fired indiscriminately into the populated al-Hawban district, north-east of Taiz, on 22 May 2017, killing three civilians, including two children, and injuring two others.

Independent Yemeni civil-society monitoring documented continued ground attacks involving government-side forces during the period surrounding the April 2022 truce. During September 2020–September 2021, Mwatana for Human Rights documented at least 110 ground attacks, killing 105 civilians, including 44 children and seventeen women, and injuring at least 311 others, including 124 children and 43 women. It attributed 71 attacks to the Houthis, eighteen to coalition or coalition-backed forces, and ten to Hadi government forces. These casualty totals concern all documented attacks and cannot be attributed exclusively to government forces. During September 2021–August 2022, Mwatana documented 59 ground attacks, killing 39 civilians, principally in Taiz, Marib, and Sa’dah. It attributed 32 attacks to the Houthis, six to government forces, five to STC forces, four to UAE-backed Joint Forces and eleven to Saudi border guards. The latter attacks affected Yemeni civilians and African migrants in border districts of Sa’dah, reportedly killing two children and injuring sixteen civilians. The reporting period extends beyond the April 2022 truce.

Historical findings: the Saudi-led coalition’s air campaign (2015–2022). The record concerning the Saudi-led coalition’s air campaign is substantially more extensive. Since March 2015, UN investigations, Human Rights Watch, Amnesty International, and Yemeni civil-society organizations have documented numerous attacks causing civilian deaths and injuries, including strikes on residential areas, markets, public gatherings, medical facilities, and other civilian sites. These investigations repeatedly raised concerns regarding distinction, proportionality, target verification, and feasible precautions. They concern external military forces operating in support of the Yemeni government and cannot be directly attributed to Yemeni government aircraft or the IRG military command.

In July 2015, Amnesty International reported having investigated eight coalition airstrikes that killed 54 civilians, including 27 children and sixteen women, and injured 55 others. Its investigations included attacks on residential areas involving large explosive weapons. Amnesty subsequently documented further apparently unlawful coalition strikes, including attacks on civilian homes and gatherings. By March 2018, it reported having investigated 36 apparently unlawful coalition airstrikes resulting in 513 civilian deaths, including at least 157 children, and 379 injuries. These figures represent cumulative investigated caseloads at different points in time and should neither be added together nor treated as comprehensive national casualty totals. Among the most serious documented incidents was the coalition’s 8 October 2016 strike on a funeral gathering in Sana’a. Human Rights Watch reported at least 100 people killed and more than 500 injured and considered the attack an apparent war crime. Its investigation identified remnants of a US-manufactured GBU-12 Paveway II laser-guided bomb.

The coalition established the Joint Incidents Assessment Team (JIAT) in 2016 to examine alleged incidents during its military operations. In its 2018 report, Hiding Behind the Coalition, Human Rights Watch identified deficiencies in its independence, impartiality, transparency, and legal analysis. These historical findings bear on the credibility of coalition assurances concerning civilian protection; investigations and remedies are considered further under ‘Accountability Mechanisms and Transitional Justice’ below.

The 2021 GEE final report (A/HRC/48/20, para. 23) continued to document civilian harm associated with coalition air operations, including the 12 July 2020 strike on a house in Beit Qateeb, Hajjah Governorate, which killed seven children and two women. It reiterated concerns regarding coalition targeting practices and the inadequacy of investigations into alleged violations. The Group’s mandate ended in October 2021, leaving an important gap in the continuation of similarly detailed independent international investigations. The UN Panel of Experts’ 2021 report (S/2021/79, para. 121 and annex 29) separately investigated coalition airstrikes on 14 February and 12 July 2020, which together killed 41 people and injured 24, mostly women and children. The July incident overlaps with the GEE investigation and should not be counted as an additional attack.

Serious concerns persisted immediately before the April 2022 truce. In a joint investigation published in April 2022, Human Rights Watch and Mwatana examined three coalition attacks in January 2022 that reportedly killed at least 80 civilians and injured 156. One strike hit a detention facility in Sa’dah, causing extensive casualties among detainees. The organizations challenged the coalition’s account of the target’s military character and examined evidence of the weaponry employed.

Continuing civilian harm and monitoring after the 2022 truce. The April 2022 UN-brokered truce substantially reduced major IRG-Houthi hostilities, although intermittent clashes, shelling and other attacks affecting civilians continued. The NCIAVHR’s successive investigations provide evidence of civilian deaths and injuries attributed jointly to government forces and coalition operations over several reporting cycles; 151 civilians killed and 123 injured in August 2020–July 2021 (Ninth Periodic Report, p. 19) ; 151 killed and 134 injured in July 2021–July 2022 (Tenth Periodic Report, p. 19); 55 killed and 96 injured in August 2022–July 2023 (Eleventh Periodic Report, p. 28); 100 killed and 62 injured in August 2023–July 2024 (Twelfth Periodic Report, p. 28); and 48 killed and 74 injured in August 2024–July 2025 (Thirteenth Periodic Report, p. 22). These figures concern combined government-force and coalition categories, not casualties attributable exclusively to regular IRG forces.

The UN Panel of Experts continued to document harm affecting civilians and civilian objects. Its 2024 report (S/2024/731, para. 135) recorded accounts of at least 153 fatalities and 180 injuries during the reporting period concerning reported Houthi attacks using artillery, rocket-propelled grenades, drones, and sniper fire. Taiz, al-Dali, Marib, al-Hudaydah and al-Bayda were particularly affected in the first half of 2024. Its 2025 report (S/2025/650, para. 60), in its discussion of frontline fighting around the Taiz axis, recorded 35 Houthi attacks on civilians and civilian objects between 1 January and 22 July 2025, killing 8 civilians and injuring 29. These figures provide context for the civilian risks in areas of government-Houthi hostilities; they do not establish IRG violations. US, UK and Israeli operations against Houthi-controlled territory likewise require separate perpetrator and operational attribution.

Renewed operations and civilian harm in 2026. The renewed escalation from July 2026 materially changed the compliance picture. After several years of comparatively limited government air operations, Yemeni government aircraft reportedly resumed offensive strikes against Houthi military targets, while Saudi aircraft subsequently conducted separate but operationally related attacks in support of the government-side offensive.

On 13 July, Yemeni government forces operating in coalition with Saudi Arabia reportedly attacked Sana’a International Airport. In its 28 July investigation, Human Rights Watch analysed satellite imagery indicating that at least five munitions were used, three of which struck the airport’s only functioning runway. According to the Yemeni government, the operation was intended to prevent an aircraft operated by the Iranian carrier Mahan Air from landing. The airport provides a critical international connection for the population of Houthi-controlled territory and supports humanitarian operations. Human Rights Watch raised serious concerns about attacks affecting civilian infrastructure, although its investigation did not establish conclusively that the particular strike violated the rules of distinction or proportionality or precautions in attack.

Escalation intensified in September. Government forces reportedly designated an extensive operational area around Mokha, Dhubab, and the Taiz–Mokha road as a ‘kill zone’, warning civilians to avoid Houthi positions, equipment and movements. Government aircraft subsequently reportedly struck Houthi military positions and vehicles within the designated area. In his 24 September 2026 statement, the UN High Commissioner for Human Rights recalled that the designation of so-called ‘kill zones’ and the issuance of warnings do not remove the protection afforded to civilians and civilian objects. Advance warnings may constitute a relevant precaution, but they do not render civilians remaining in the area lawful targets, convert civilian objects into military objectives, or dispense with the other IHL rules on targeting in each individual attack.

The renewed fighting caused many civilian casualties. On 10 September, the UN Special Envoy reported that civilians had been killed in Taiz, al-Hudaydah, al-Bayda, al-Jawf and elsewhere, including children reportedly killed by shelling of displacement sites in Marib. On 15 September, UN Assistant Secretary-General reported that OHCHR had recorded forty civilian casualties, including eight deaths and 32 injuries, during the intensified fighting. The circumstances of these deaths were not sufficiently described to permit an assessment of whether IHL had been violated.

At the UN Geneva briefing on 18 September 2026, OHCHR reported 110 civilian casualties in 31 incidents since hostilities escalated in August: 28 killed and 82 injured. It warned that insecurity and access constraints impeded verification. In its 24 September statement, OHCHR documented at least 125 civilians killed or injured since early August, including 32 killed, among them fifteen children and three women. These casualty totals do not provide a complete perpetrator breakdown permitting a separate total for casualties caused by Yemeni government or Saudi air operations.

But displacement also increased sharply: according to the International Organization for Migration, 5 October 2026, more than 184,000 people had been displaced between 1 August and 4 October. By 7 October, IOM reported that displacement had exceeded 200,000 people.

New reports of civilian harm followed. Associated Press reported on 11 October that the Houthi-run health ministry said a suspected Saudi strike on a house in Sana’a killed a child and injured six people; Saudi authorities had not commented. AP’s 12 October report relayed Houthi health-authority claims that strikes on a government complex and a house in al-Jawf killed nine people, including five children and a woman, and injured seven. These are attributed allegations requiring investigation; the reporting does not independently establish the responsible aircraft, targets or legality of the attacks.

Available evidence therefore establishes a historical record of indiscriminate ground attacks involving government-side forces and extensive civilian harm from Saudi-led coalition air operations, including incidents involving identifiable transferred munitions. Subsequent investigations by NCIAVHR and Mwatana demonstrate that civilian harm involving government forces continued during the period of reduced hostilities. The renewed use of Yemeni and Saudi air power in 2026 raises additional concerns, particularly in populated areas and near civilian infrastructure. However, available independent reporting does not yet provide a comprehensive actor-by-actor breakdown of civilian casualties or sufficient incident-specific evidence to assess systematically the legality of individual strikes. These limitations do not establish compliance, but they preclude attributing aggregate civilian casualties or specific violations to the IRG without further corroboration.

Treatment of persons in the power of the enemy

UN investigations and independent human rights reporting have documented serious violations involving the treatment of persons captured or detained in connection with Yemen’s armed conflicts, including arbitrary detention, enforced disappearance, torture, ill-treatment and deaths in custody. These practices have implicated Yemeni government military and intelligence structures, government-aligned armed formations and foreign military partners. However, the fragmentation of government-side forces requires careful differentiation between violations attributable to the IRG regular armed forces, those committed by affiliated armed groups operating under separate command structures, and those involving security or intelligence agencies exercising detention powers.

Common Article 3 of the Geneva Conventions prohibits murder, cruel treatment, torture, hostage-taking and outrages upon personal dignity against persons taking no active part in hostilities, including members of armed forces who have laid down their arms or been placed hors de combat by detention. Additional Protocol II establishes further protections for persons deprived of liberty for reasons related to the armed conflict, including humane treatment, adequate conditions of detention, and fundamental judicial guarantees. The present assessment focuses on conflict-related detention involving military forces and organized armed groups. Relevant conduct by intelligence or security agencies is included where it directly concerns the treatment of persons held because of their actual or perceived affiliation with an opposing party; broader law-enforcement detention practices are examined separately, below.

Historical findings concerning government military forces and affiliated armed formations. The GEE, in its 2019 detailed findings (A/HRC/42/CRP.1, paras. 221–253), documented arbitrary detention, enforced disappearance, torture and other ill-treatment involving Yemeni government forces, UAE-backed formations and foreign military personnel in Aden, Shabwah and Hadramaut. It identified detention without judicial oversight, denial of access to relatives and legal representation, and failures to comply with release orders. The findings included al-Tin, a detention facility controlled by Yemeni armed forces and Saudi forces, where the Group documented prolonged detention without proceedings and torture during interrogation (paras. 240–243 and 248–253). It separately identified torture and sexual violence at UAE-controlled al-Bureiqa and Security Belt-controlled Bir Ahmed II. These findings distinguish direct involvement by Yemeni military structures from abuses by foreign personnel and separately commanded formations; they do not attribute every incident to regular Ministry of Defence forces.

The GEE’s 2020 detailed findings (A/HRC/45/CRP.7, paras. 161–65) provided additional evidence concerning government-associated military formations. In Taiz, the GEE verified eight cases of enforced disappearance involving the Taiz Axis Command and Islah-affiliated actors, with disappearances lasting from several months to more than four years. It documented allegations of torture and other ill-treatment, including sexual violence, and identified public buildings believed to have been used as unofficial detention sites, including al-Nahda school, the Public Funds Office, the Oversight and Accounting Bureau, and the National Institute. The 22nd Mechanised Brigade, military intelligence, and government-aligned resistance formations were among the actors implicated in the reported disappearances. The GEE also examined the cases of Ayoub Salihi and Akram Hameed, both of whom had remained disappeared for more than four years. These findings are particularly relevant because they identify military formations and intelligence structures associated with the Taiz Axis, rather than merely referring to violations occurring in government-controlled territory.

The GEE (paras. 164–165 and 329) also identified serious shortcomings in the response of military and judicial authorities. Correspondence reviewed by the Group showed that prosecutors had requested information about disappeared persons from military and security authorities in Taiz. Although five cases were reportedly referred to competent authorities in October 2019 and the individuals subsequently released, the GEE identified a general lack of effective investigations. It also reported intimidation of activists critical of military conduct, including summons issued by the Taiz Axis Command.

The GEE’s 2021 report (A/HRC/48/20, paras. 40–42), covering July 2020–June 2021, documented further allegations of arbitrary detention and enforced disappearance involving government and affiliated forces. The cases involving medical workers, government-associated detention and STC counterterrorism forces are presented under Arbitrary detention and enforced disappearance in the law-enforcement assessment, below. They remain relevant here where detention was connected with the armed conflict.

Mwatana for Human Rights’ 2020 investigation, In the Darkness, provides particularly detailed evidence of detention-related abuses by identifiable government military structures, foreign military personnel and affiliated armed formations. Based on more than 2,500 interviews and investigations of eleven unofficial detention sites, it documented abuses across multiple institutions. Its mixed military/security detention and disappearance figures are presented under Arbitrary detention and enforced disappearance in the law-enforcement assessment below; torture and custodial-death figures appear under Torture, ill-treatment and sexual violence. The facility-specific findings discussed here identify the military structures and allied formations involved.

In Marib, Mwatana identified detention facilities controlled by Islah-affiliated government forces loyal to President Hadi, including the Military Intelligence Division prison of the Third Military Region and the separate Political Security Department. Civilians were reportedly detained alongside combatants, with some arrests occurring at the al-Falaj checkpoint and apparently based on perceived Houthi affiliation, family surnames, or northern origin. Former detainees described beatings with cables, overcrowding, inadequate ventilation, and severe restrictions on water and sanitation. One detainee described approximately twelve prisoners held in a cell measuring roughly three by three metres. The military intelligence facility is particularly relevant to the present assessment because of its institutional connection with the government military structure; detention and torture involving the separate Political Security Department are examined in Arbitrary detention and enforced disappearance and Torture, ill-treatment and sexual violence in the police compliance section below.

The GEE’s 2020 detailed findings (A/HRC/45/CRP.7, paras. 157–58) examined conflict-related detention at Marib Political Security Prison between 2016 and 2020. Civilians travelling from Houthi-controlled areas or returning from abroad were reportedly arrested on suspicion of Houthi affiliation, sometimes because of their surnames or northern origin, and held incommunicado without charges or judicial proceedings for periods extending to several years. Although Political Security is an intelligence institution rather than a regular Ministry of Defence unit, the cases are relevant here because they concern persons detained on the basis of perceived affiliation with an opposing party to the conflict.

The GEE’s 2020 detailed findings (para. 168) verified torture, including sexual violence, at Marib Political Security Prison and the linked Al-Saleh Institute facility. Its investigation examined the treatment of five men and two boys. Political Security personnel used beatings, electric shocks, suspension, mock executions, and violence targeting detainees’ genitals to humiliate them and compel admissions of alleged Houthi affiliation. One boy suffered lasting injuries following suspension and assaults during interrogation. These findings concern interrogation over alleged affiliation with an opposing party and are retained here as conflict-related detention evidence.

UN Security Council monitoring and national investigations. The UN Panel’s 2021 report (S/2021/79, paras. 128–131) examined detentionby government forces; its detention figures are presented under Arbitrary detention and enforced disappearance, below. It also investigated two extrajudicial killings allegedly committed by government militarypersonnel inside hospitals in Taiz, in March 2019 at al-Tharwah Hospital and November 2020 at Rawdah Hospital. The Panel documented an arbitrary detention, an assault, and threats against members of the judiciary by elements associated with the government military, apparently intended to obstruct investigations or prosecutions. The subsequent 2022 Panel report (S/2022/50, paras. 96 and 105–06) documented further arbitrary arrests involving government and STC-affiliated forces and weaknesses in judicial oversight, including failure to implement judicial decisions.

Recent developments concerning conflict-related detainees. In May 2026, the IRG and the Houthis reportedly reached an agreement on a large-scale exchange involving more than 1,600 conflict-related detainees. According to Reuters, 14 May 2026, the agreement envisaged the release of approximately 1,100 detainees by the Government and 580 by the Houthis, including Saudi and Sudanese nationals. The agreement builds on the December 2018 Stockholm Agreement, which established a mechanism for exchanging prisoners, detainees, missing persons, and persons arbitrarily detained or forcibly disappeared in connection with the conflict. The Office of the UN Special Envoy and the ICRC co-chair the Supervisory Committee responsible for facilitating the exchange process. Subsequent agreements and releases occurred in 2020 and 2023, although the broader detention issue remained unresolved.

Continuing detention-related violations and monitoring. The NCIAVHR’s Fourteenth Periodic Report (August 2025–July 2026, pp. 58 and 69–70) continues to identify detention-related violations involving government military and security bodies. Its aggregate categories do not establish the precise detaining authority or an armed-conflict nexus in every case. The UN Panel of Experts also reported in its October 2025 report (S/2025/650, para. 137) that it had documented arbitrary detention, including of children, in government-controlled areas, notably Aden, Shabwah, and Lahj, involving individuals from northern regions suspected of Houthi affiliation. The Government denied violations of the rule of law.

Treatment of captured, surrendering and other persons hors de combat. Beyond abuses in detention facilities, the treatment of opposing fighters during military operations raises separate concerns under Common Article 3 and customary IHL. Persons who have surrendered, been captured or otherwise become hors de combat must not be subjected to violence, including deliberate killing, torture or cruel treatment. The prohibition applies irrespective of whether they are members of an opposing armed group and independently of their prior conduct in hostilities. The Panel’s investigation of alleged killings by government military personnel inside hospitals in Taiz is relevant, although the available summary does not establish whether the victims werehors de combatat the time. Detention-related abuses should likewise be distinguished from incidents involving the killing or mistreatment of fighters at the point of capture or surrender.

A serious recent allegation concerns the treatment of Houthi fighters during operations near the Bab el-Mandeb Strait in October 2026. On 6 October, Reuters reported that it had verified videos showing armoured vehicles driving over Houthi fighters during operations to retake territory near the strait. Some individuals were armed or moving, while others appeared unarmed, stationary or with their hands raised. The Giants Brigades, identified as the principal government-aligned formation involved in the fighting, announced an investigation; its spokesperson stated that the brigade believed some opposing fighters were surrendering before others opened fire. If individuals were surrendering, incapacitated or otherwise hors de combat when deliberately struck, such conduct would constitute a serious violation of IHL and a likely war crime. The footage raises serious concerns requiring independent investigation, particularly regarding the status of the individuals targeted, the circumstances of the incident and the responsibility of the personnel and commanders involved. The identity of the personnel involved and the Giants Brigades’ legal status at the time require investigation. Their position outside the regular MoD chain and their external support arrangements do not, by themselves, exclude attribution to Yemen; State responsibility must be distinguished from the individual criminal responsibility of perpetrators and commanders.

Child recruitment and use

The recruitment and use of children by the IRG armed forces and affiliated armed formations is a recurring concern. Yemen is a party to the Optional Protocol to the Convention on the Rights of the Child on the involvement of children in armed conflict (OPAC), which prohibits compulsory recruitment of persons under eighteen into State armed forces and requires feasible measures to prevent members under eighteen from taking a direct part in hostilities. Article 4 prohibits recruitment or use in hostilities of persons under eighteen by armed groups distinct from State armed forces. Under Article 4(3)(c) of Additional Protocol II, children under fifteen must neither be recruited into armed forces or groups nor allowed to take part in hostilities. Conscripting or enlisting children under fifteen into armed forces or groups, or using them to participate actively in hostilities, constitutes a war crime under customary international law. Yemen’s declaration upon ratification of OPAC confirms eighteen as the minimum age for voluntary recruitment into its armed forces, while, as reported by the Sana’a Center, Article 149 of Law No. 45 of 2002 on Child Rights prohibits the recruitment of persons under eighteen and their involvement in armed conflict.

Mwatana’s 2020 submission to the UN Committee on Economic, Social and Cultural Rights reported documenting at least 1,117 cases of child recruitment or use for military purposes in 2018, attributing approximately 11% to President Hadi-aligned forces and popular resistance groups and 17% to UAE-backed Security Belt and Hadhrami Elite forces. Mwatana’s investigation, Colored Coffins: Recruitment and Use of Child Soldiers in Armed Conflict in Yemen, April 2013–December 2018, examined recruitment practices, children’s military assignments, and the consequences for recruited children.

Mwatana’s February 2024 briefing, Fragile Bodies in the Heat of War, documented 1,238 incidents of child recruitment and use between 2014 and December 2023, including 212 attributed to IRG forces, 127 to the STC, 18 to the Joint Forces, and 20 to the Saudi/UAE-led coalition. The briefing also documented the recruitment and military training of a child by the Government’s 101st Infantry Brigade in Al-Jawf, followed by deployment to frontlines against the Houthis. According to testimonies, the child sustained serious injuries when a landmine struck a military vehicle in June 2022 and died three months later.

The GEE’s 2020 detailed findings (A/HRC/45/CRP.7, paras. 247–53) provide particularly relevant evidence of recruitment by government-associated forces. Between May 2016 and January 2020, the GEE documented boys aged twelve to seventeen recruited in Taiz and Lahj and transferred to Saudi Arabia for military training before being used in combat by brigades allegedly connected with coalition and/or Yemeni government personnel. The Group reviewed documents suggesting possible links between these formations and the Saudi-led coalition’s Joint Forces Command and received allegations of coalition training, financing and material support. However, it expressly stated that further investigation was necessary to establish the formations’ command and control arrangements.

Mayyun Organization for Human Rights, in its Child Warriors report, documented the recruitment and use of 2,233 children by parties to the conflict between July 2021 and December 2022. It attributed 24 cases (1.1%) to the IRG and affiliated armed formations, compared with 2,209 (98.9%) to the Houthis.

In 2014, the Government signed a UN action plan to end and prevent recruitment and use of children by its armed forces, followed by an implementation road map in 2018. It subsequently adopted directives prohibiting child recruitment and measures concerning age verification, military inspections, and child-protection structures. The UN Secretary-General’s report on children and armed conflict in Yemen (S/2025/113, paras. 13–17), covering January 2021–December 2023, verified 564 cases of recruitment and use of children aged eight to seventeen: 259 in 2021, 132 in 2022, and 173 in 2023. Of the 564 children recruited and used by all parties during 2021–2023, 293 were assigned combat roles, including 27 children under fifteen. The report acknowledges access and security restrictions and communities’ fear of retaliation when reporting violations (paras. 3 and 15).

The Government’s action plan and implementation measures are documented in S/2025/113, paras. 51–57. The Secretary-General’s 2022 annual report (A/76/871–S/2022/493, para. 305) delisted government forces, including the Yemen Armed Forces, for recruitment and use following progress in implementation and a significant decrease in verified cases. The decision was expressly conditional on completion of pending action-plan activities and a continued decrease in recruitment, with twelve months of UN monitoring and the possibility of relisting. It did not establish that recruitment had ceased permanently or extend automatically to every government-aligned formation; pro-government militias and the Security Belt Forces remained separately listed (S/2025/113, para. 2).

The UN Secretary-General’s report covering 2025 (A/80/723–S/2026/357, para. 233) verified the recruitment and use of 262 children (259 boys and three girls), including children serving in combat and support roles. Of these, 135 were attributed to the Yemen Armed Forces and affiliated forces and groups, compared with 127 attributed to the Houthis. The government category included 29 children attributed to the Yemen Armed Forces, sixteen to the National Shield Forces, twelve to the 5th Support and Reinforcement Brigade, ten to the Shabwani Defence Forces, seven to the Security Belt Forces and three to the Southern Transitional Council, among other formations. 

Attacks against civilian objects, including schools and medical facilities

The protection of civilian objects has been a recurring concern in the conduct of hostilities by Yemen’s IRG, government-affiliated armed formations and the Saudi/UAE-led coalition. The NCIAVHR, in its Thirteenth Periodic Report (pp. 26–27), documented two artillery attacks affecting residential areas in Al-Bayda Governorate, attributing both to the Security Belt Forces under Abdul Aziz Al-Mansouri, Commander of the Yafa’a Command. On 2 June 2022, artillery fire struck a family home in Dhi Makhshab village, Al-Zahir district, partially destroying the building, killing a five-year-old girl and injuring her infant brother and mother. The Commission found no evidence of armed personnel or military barracks in the village, although clashes were occurring approximately 400 metres away. On 24 October 2022, artillery fire struck the centre of Ainah village in the same district, among civilian homes, injuring three children. Witnesses reported no military positions within the village, and the Commission attributed responsibility to the same Security Belt command.

Historical GEE investigations documented coalition airstrikes affecting civilian and protected objects. Its 2019 detailed findings (A/HRC/42/CRP.1, paras. 186–193) examined attacks on a residential building, a cattle market, a school and a mosque in Aden and Lahj. Its 2020 detailed findings (A/HRC/45/CRP.7, paras. 66–74) investigated further strikes affecting homes and a detention facility and raised concerns about target verification, proportionality and precautions. These findings are relevant to the risks associated with aircraft and aerial munitions supplied to external forces supporting the IRG. They should retain their coalition attribution; they do not establish that Yemen Armed Forces personnel carried out the strikes or that every coalition operation is legally attributable to Yemen.

Schools and educational facilities. Mwatana’s 2020 investigation, Undermining the Future, provides substantial historical evidence of government involvement in attacks affecting educational facilities. Covering March 2015–December 2019, it documented more than 380 incidents involving schools, including 36 attacks during ground fighting. Eight were attributed to forces loyal to the internationally recognised government, while six involved both government and Houthi forces. Of 171 documented instances of military use or occupation of schools, thirty were attributed to government and affiliated Popular Resistance forces and eight to UAE-backed STC forces. Separately, Mwatana attributed 153 airstrikes affecting schools to the Saudi/UAE-led coalition.

The Secretary-General’s report on children and armed conflict in Yemen covering 2019–2020 (S/2021/761, paras. 33–36) verified 37 attacks on schools, including three attributed to the Yemeni armed forces and two crossfire incidents: one involving the armed forces and the Houthis, and one involving the armed forces and Popular Resistance (para. 33).

Hospitals, medical facilities and personnel. Medical facilities and personnel have also been affected by attacks, armed occupation, and interference with healthcare. The joint investigation by Mwatana and Physicians for Human Rights, I Ripped the IV Out of My Arm and Started Running (2020), documented 120 incidents affecting healthcare facilities and personnel between March 2015 and December 2018. These included aerial and ground attacks, military occupation, assaults on medical personnel and interference with healthcare services. The investigation attributed violations to the Saudi/UAE-led coalition, government forces and the Houthis, while distinguishing the coalition’s aerial attacks from the use of indirect-fire weapons, occupation of facilities, and intimidation of medical staff by other parties.

Mwatana’s 2020 submission to the UN Human Rights Council provides more specific government-attributed evidence. Between September 2019 and August 2020, the organisation verified 22 incidents affecting healthcare facilities and personnel, attributing thirteen to forces loyal to the internationally recognised government, eight to the Houthis and one to an unidentified perpetrator. The documented conduct included armed incursions into medical facilities, threats or assaults against healthcare workers and ground attacks affecting medical premises. The NCIAVHR’s Tenth Periodic Report (pp. 10 and 33) attributed four of seventeen investigated incidents affecting medical facilities to the Army of the IRG.

The Secretary-General’s report on Children and armed conflict covering 2021–2023 (S/2025/113, paras. 42–44) verified 62 attacks affecting hospitals and protected persons associated with hospitals, attributing 32 to the Yemen Armed Forces and affiliated forces and groups. Of these, thirteen were attributed directly to the Yemen Armed Forces, eleven to the Security Belt Forces, two each to the Amaliqah Brigades, STC and Yemeni police, one to the Thunderbolt Forces and one jointly to the Yemen Armed Forces and Shabwani Elite Forces. The incidents included damage to medical buildings and equipment, intimidation and assault of medical personnel and patients, interference with medical services, abductions, and looting. The UN’s classification therefore encompasses different forms of attacks and violence affecting healthcare, not exclusively bombardment or the deliberate targeting of hospital buildings.

Conflict-related sexual violence

Rape and other forms of conflict-related sexual violence have been documented among violations committed by Yemen’s armed and security forces and other parties to the conflict. Women, girls, men, and boys have been affected, with particular risks documented among displaced populations, migrants, children, and marginalised communities. Rape and other forms of sexual violence are prohibited under international humanitarian law, including Common Article 3 of the Geneva Conventions and Article 4(2)(e) of Additional Protocol II, and may constitute war crimes. The following findings focus on violations attributed to the Yemen Armed Forces and government-affiliated formations, while distinguishing the conduct of separately commanded UAE-backed forces and other parties.

The UN GEE, in its 2019 detailed findings (A/HRC/42/CRP.1, paras. 295–299), investigated allegations of sexual violence by members of the 35th Armoured Brigade of the Yemen Armed Forces in Al-Shamayatayn district, Taiz Governorate, between June 2017 and 2019. The Group documented seven cases involving rape, attempted rape and other sexual assaults against members of marginalised communities, including women, a man, boys and a girl. It identified the involvement of multiple armed perpetrators, military vehicles and, in six of the seven cases, personnel acting in the course of their purported duties. The violations included gang rape, and sexual violence used to intimidate or humiliate victims and their communities. The GEE found reasonable grounds to believe that members of the brigade had committed rape and other forms of sexual violence potentially amounting to war crimes.

The GEE also documented sexual violence by UAE-backed Security Belt Forces against migrants and members of marginalised communities in Aden. Its 2019 findings (A/HRC/42/CRP.1, paras. 290–294) and 2020 report (A/HRC/45/6, para. 69) included rape, sexual assault and forced nudity, with weapons and threats used to intimidate survivors. The detailed findings and victim figures appear under ‘Torture, ill-treatment and sexual violence in the law-enforcement assessment’below. They are the same incidents, relevant here where connected with armed conflict. UAE backing and distinct command arrangements do not alone establish or exclude attribution to Yemen; the formation’s legal status and its relationship with Yemeni and Emirati authorities at the time require separate assessment.

More recent UN monitoring establishes continued involvement of regular government forces and affiliated formations in sexual violence against children. The Secretary-General’s report on Children and armed conflict covering 2021–2023 (S/2025/113, para. 31) verified sexual violence against 46 children, comprising 22 boys and 24 girls aged five to seventeen. Thirty violations were attributed to the Yemen Armed Forces and affiliated forces and groups, including sixteen directly to the Yemen Armed Forces, six to Yemeni police, two each to the Amaliqah Brigades and 5th Brigade, and one each to the Special Forces Brigade, Security Belt Forces, Shabwani Defence Forces, and Hadrami Elite Forces. One further violation was jointly attributed to the Yemen Armed Forces and Security Belt Forces. The UN reported that some children had been raped repeatedly and threatened to prevent disclosure or access to assistance. This is a mixed actor category: the six police-attributed violations are distinguished from the sixteen attributed directly to the Yemen Armed Forces. Inclusion under this thematic account does not by itself establish the operational function or armed-conflict nexus of every incident.

The Secretary-General’s annual report on Children and armed conflict covering 2025 (S/2026/357, para. 236) verified rape against twelve children, including two cases involving gang rape. Ten violations were attributed to the Yemen Armed Forces and affiliated forces and groups, including three directly to the Yemen Armed Forces, two to the STC and one to the Security Belt Forces; two were attributed to the Houthis. The UN also reported that one boy was killed following rape. These findings demonstrate the persistence of grave sexual violence against children involving IRG and affiliated forces during 2025.

Conflict-related sexual violence remains substantially underreported, owing to stigma, fear of retaliation, obstacles to accessing assistance, and the particular vulnerability of displaced and marginalised populations. The findings nevertheless identify violations involving regular Yemen Armed Forces personnel, police and affiliated armed formations across successive reporting periods. For the purposes of assessing arms transfers, the documented involvement of armed personnel, the use of weapons and military vehicles in some historical incidents, and the recurrence of violations against children raise concerns regarding discipline, command oversight, prevention, and accountability. These concerns require differentiation between forces formally incorporated into the IRG’s command structure and formations whose effective control remains disputed.

Landmines and cluster munitions

Landmines and cluster munitions have caused extensive civilian harm in Yemen, both during hostilities and through continuing explosive contamination. Responsibility for their use must, however, be distinguished between Yemeni government forces, the Saudi/UAE-led coalition and the Houthis. Yemen has been a State Party to the Anti-Personnel Mine Ban Convention since 1 March 1999. It is not a party to the Convention on Cluster Munitions, although the use of cluster munitions remains subject to applicable IHL rules on distinction, indiscriminate attacks, proportionality and precautions in attack.

Documented anti-personnel mine use by Yemeni government forces predates the current armed conflicts. According to the Landmine Monitor, government forces laid GYATA-64, PMN-1, and PMD-6 anti-personnel mines around military positions in Bani Jarmooz in 2011–2012. In November 2013, the Yemeni Prime Minister’s office acknowledged that their use violated Yemen’s treaty obligations, and a military investigation was subsequently initiated. This establishes a historical treaty violation involving government forces but does not demonstrate continued use by the internationally recognised government’s (IRG) armed forces since 2015.

Since 2015, the mine incidents documented in the sources reviewed have been attributed predominantly to the Houthis. Mwatana’s report, Fields of Death: Civilian Victims of Landmine Incidents 2016–2024, documented 537 landmine incidents across fourteen governorates between January 2016 and March 2024, attributing 512 to Ansar Allah and 25 to other parties. The report covers anti-personnel and anti-vehicle mines, so these figures should not be presented as an anti-personnel-only total. Nor does the unspecified ‘other parties’ category establish responsibility by particular IRG formations. The GEE’s 2019 detailed findings (A/HRC/42/CRP.1, paras. 202–209, 336–344 and 570–576) separately investigated mine incidents and found reasonable grounds to believe that Houthi or Houthi-Saleh forces had used anti-personnel and anti-vehicle mines unlawfully in Aden/Lahj, Taiz and Hudaydah.

The UN Panel’s 2024 report (S/2024/731, para. 134 and annex 187) records government officials’ information that at least 61 civilians were killed and 119 injured by landmines, sea mines and other remnants of war in the first half of 2024. It also describes Houthi local mine manufacture and triggering mechanisms increasing risks to civilians and deminers. Separately, the UNMAS 2024 Annual Report reports 93 casualties recorded by UNMHA in Hudaydah during 2024, more than forty per cent involving women and children. The datasets cover different geographic areas, periods and explosive hazards and should not be combined. Neither casualty totals nor clearance of old mines establishes when mines were laid or identifies the force responsible for each incident. These particular sources do not establish new anti-personnel mine use by regular IRG forces or coalition personnel.

Although Yemen is not bound by the Convention on Cluster Munitions, their use, particularly in populated areas, raises serious concerns under customary IHL because of their wide-area effects and unexploded submunitions. Yemen historically acquired cluster munitions and may retain stocks. According to the Cluster Munition Monitor’s 2026 Yemen profile, Jane’s reported KMGU submunition dispensers in Yemeni Air Force service in 2004, alongside Uragan multiple-launch rocket systems and aircraft capable of delivering Soviet-designed RBK cluster bombs. These historical holdings are relevant to assessing potential availability within Yemen’s military inventory but do not establish their current operational status, custody or use by IRG forces.

The Saudi-led coalition used air-delivered and ground-launched cluster munitions during operations against Houthi forces between 2015 and 2017. According to the Cluster Munition Monitor 2026, at least 23 attacks were documented during this period, involving seven identified categories of cluster munitions and submunitions. These included US-manufactured CBU-105 Sensor Fuzed Weapons, CBU-87, and CBU-58 bombs, UK-manufactured BL755 bombs, Brazilian ASTROS II rockets, and US-manufactured M26 rockets, as well as ZP-39 submunitions of unidentified origin and delivery system. Human Rights Watch documented civilian casualties and unexploded submunitions associated with coalition attacks, including CBU-105 use in 2015. These findings contributed to scrutiny of US cluster munition transfers to Saudi Arabia, which were suspended in May 2016. The coalition acknowledged using CBU-105 weapons and, in December 2016, UK-manufactured BL755 bombs, subsequently committing to cease use of the latter. The Monitor has found no compelling evidence of further coalition cluster munition use in Yemen after 2017.

Unexploded submunitions continue to endanger civilians long after cluster munition use. According to the Landmine and Cluster Munition Monitor, cluster munition remnants caused 30 recorded casualties in 2023 and one in 2024. Cluster Munition Monitor 2026 continues to identify Yemen as substantially contaminated, while Mine Action Review highlights the fragmentation of mine-action institutions and difficulties conducting survey and clearance operations across territories controlled by different authorities. These continuing casualties and clearance challenges reflect the consequences of past use and cannot be interpreted as evidence of new cluster munition attacks.

Accountability Mechanisms and Transitional Justice

Accountability for serious violations of IHL and IHRL remains limited in Yemen, including for violations attributed to the IRG, government-affiliated formations, and external forces supporting it. Domestic and coalition mechanisms have documented or assessed allegations, but their existence must be distinguished from effective criminal investigations, prosecutions, and remedies. Human Rights Watch’s World Report 2026 reports that accountability for violations committed by parties to the conflict remains largely absent.

The principal national investigative mechanism is the NCIAVHR, established by Republican Resolution No. 140 of 2012 and subsequent amendments. Its mandate covers alleged violations committed by all parties throughout Yemen. In presenting its Fourteenth Periodic Report on 17 September 2026, the Commission described continuing field investigations, interviews, evidence preservation, and visits to official and unofficial detention facilities.

The NCIAVHR’s Fourteenth Periodic Report (p. 14) documented 2,062 incidents affecting 2,932 victims across Yemen. These covered more than 36 types of human rights violations. Its cumulative documentation since it began work in January 2016 reached 35,015 incidents involving 70,470 victims. These totals encompass different parties and categories of violations; they are neither a measure of IRG responsibility alone nor a record of criminal convictions. The Commission’s official launch summary also acknowledged restricted access to victims and witnesses, weak enforcement of investigative outcomes and disrupted avenues for remedy and reparation, and called on the government to expedite investigations and prosecutions and strengthen judicial independence. 

Earlier UN findings identified structural obstacles to translating documentation into accountability. In its 2021 report (A/HRC/48/20, para. 75), the GEE highlighted compromised judicial independence, politicisation, inadequate protection of victims and witnesses, violations of fair-trial rights, and gender bias. These historical findings remain relevant when assessing the capacity of domestic institutions to investigate alleged violations by military and security personnel, although they should not be presented as a current inventory of pending proceedings.

Coalition accountability follows a separate institutional process. As noted in the discussion of coalition airstrikes above, the Joint Incidents Assessment Team (JIAT) assesses selected allegations concerning coalition operations. Human Rights Watch’s 2018 report, Hiding Behind the Coalition, found deficiencies in transparency, impartiality, and independence, as well as inadequate legal analysis and questionable conclusions concerning particular attacks. Coalition members retain their own obligations to investigate alleged serious violations by their forces, prosecute those responsible where warranted and provide reparation. JIAT assessments and recommendations should therefore not be equated with completed criminal proceedings or effective redress, and they do not establish accountability within the IRG’s military command structures.

The historical 2018 Panel findings on UAE-associated detention, discussed under Treatment of persons in the power of the enemy above, also require scrutiny of Yemeni authorities’ own conduct. Accountability is not limited to identifying foreign perpetrators: it includes investigating the involvement of Yemeni personnel, safeguards during detainee transfers and failures to exercise available oversight powers. The Panel’s findings do not establish that those failings remain unchanged, but the mere existence of national or coalition investigative bodies does not demonstrate that they have been remedied.

The international accountability framework also remains restricted. The Human Rights Council’s October 2021 decision not to renew the GEE’s mandate ended the principal independent international investigative mechanism dedicated to violations by all parties in Yemen. Human Rights Watch’s 2026 assessment identifies the continuing absence of an equivalent independent international mechanism. The UN Panel of Experts on Yemen nevertheless continues sanctions-related investigations. Its work can support scrutiny of responsible actors, but its mandate does not provide a substitute for judicial remedies.

Yemen is a signatory but not a State Party to the Rome Statute of the International Criminal Court. As such, the ICC does not have jurisdiction over the acts committed in the armed conflicts in the country. Saudi Arabia is not a signatory to the Rome Statute. No criminal prosecutions are believed to have been mounted for war crimes committed by the IRG Armed Forces, the Houthis, or Saudi Arabia in the course of the ongoing armed conflicts.

Key Weaponry in Law Enforcement Agencies

Legal Framework

Ordinary security in areas controlled by the IRG is formally the responsibility of the Ministry of Interior (MoI) and the Police Authority governed by Law No. 15 of 2000 on the Police Authority. Article 4 describes the police as a civilian, regular institution headed by the Minister of Interior. Article 3 also defines the regular police forces as quasi-military. Article 7 assigns responsibility for maintaining public order and security; preventing and investigating crime and apprehending offenders; protecting persons and property; securing public assemblies and processions; administering prisons and guarding prisoners; protecting public facilities; and executing judicial orders.

The formal institutional structure nevertheless does not fully correspond to the security environment in government-controlled Yemen. Research by the Yemen Policy Center describes a fragmented system in which ordinary police operate alongside military, paramilitary, and informal security providers, with competing political affiliations and unequal resources. The distinction between formal police and other forces exercising security powers remains important in Aden, Hadramawt, and Shabwah. Examples demonstrate overlapping policing and military-security functions.

As discussed, the restructuring underway since January 2026 includes a distinct process of incorporating internal-security formations into the MoI. The Sana’a Center reports that former Security Belt Forces were placed under the Ministry and renamed the National Security Forces, with other formerly STC-aligned elements incorporated into this structure. These forces operate alongside the Aden Police and Emergency Road and Patrol Police. Their institutional evolution should therefore be distinguished from the restructuring of military formations under the Ministry of Defence or other military command arrangements (Sana’a Center, Military and Security, January–March 2026).

The process extends beyond renaming. The Sana’a Center identifies the absorption of former Aden Security and Lahj Security forces into the National Security Forces, with the ultimate objective of integrating these groups into the MoI’s Special Security Forces. It also identifies integration initiatives involving personnel from the Hadrami Elite, Security Support, and Hadramawt Protection forces. These measures seek to increase personnel available for internal security, particularly in cities such as Aden and Mukalla. The Center nevertheless presents integration as an ongoing process whose durability depends partly on political loyalties and payment of salaries.

Institutional integration should consequently be understood as an evolving process. ACLED’s revised Yemen actor classifications, as of February 2026, distinguished the National Security Forces as a police-level State actor, while retaining Hadrami Elite and Shabwani Defence formations as military actors. These are analytical classifications rather than legal determinations, but they reinforce the need to distinguish personnel integrated for policing purposes from formations retaining military functions.

The Police Act regulates the use of force and firearms. Article 9 limits force during disorder to what is necessary and prohibits physical torture and psychological coercion during evidence-gathering or detention. Article 10 permits weapons or firing only as the sole means of achieving the objective, to the necessary extent, with efforts to avoid fatal injury. Its grounds include arrest involving serious crime and resistance or flight; preventing offences punishable by death or at least five years’ imprisonment; armed public-security offences; protecting guarded persons or places; self-defence; overcoming forcible resistance to lawful duties; and prison intrusion or escape. Article 11 prescribes an initial warning, a warning shot, and targeting non-fatal areas where possible.

Prime Minister’s Decision No. 35 of 2002, issuing the Implementing Regulation, elaborates these procedures. Article 8 requires exhaustion of available alternatives before firing in the circumstances it addresses, an audible warning and a warning shot where necessary. Authority normally rests with a superior, patrol commander or the most senior assigned officer, although individual action is permitted where necessity precludes obtaining an order. For public disorder, it prescribes warnings, a dispersal period and an exit direction, followed by dispersal measures and further warnings. It also authorises escalation to firing where a crowd refuses to disperse despite these measures or where serious attacks occur against personnel, persons or property. Article 9 requires the force commander to control firing through separate orders for each shot and pauses allowing dispersal or surrender.

PolicingLaw.info assesses Yemen’s framework as incompatible with international standards and states that police firearms use must be confined to confronting an imminent threat of death or serious injury, or a grave and proximate threat to life. Its profile also recognises that governmental and non-governmental forces perform law-enforcement functions. The profile therefore supports the assessment that Yemen needs tighter restrictions on firearms use.

For this assessment, ‘police’ thus denotes the formal Police Authority where the evidence concerns that institution; the broader law-enforcement assessment also encompasses other government-aligned formations when they exercise policing powers. Their institutional status, actual command relationships and function in the operation concerned should be identified separately. Renaming or administrative incorporation under the MoI provides evidence of institutional change, but does not by itself establish unified operational control or effective implementation of police use-of-force rules.

Oversight

Oversight of the formal police combines internal inspection and disciplinary procedures with judicial supervision, but PolicingLaw.info identifies no independent civilian police oversight body in Yemen. Within the MoI, the General Directorate of Inspection and Oversight is responsible for regular and unannounced inspections of subordinate bodies and governorate security directorates, including reviewing compliance with laws, examining detention premises and assessing treatment of the public.

Police misconduct may attract disciplinary sanctions without excluding criminal liability. Article 92 of the Police Act provides for reprimands, warnings, salary deductions, suspension, delayed or withheld annual increments, delayed promotion, and termination of service. Article 93 requires a written investigation, an opportunity for the officer to respond and present a defence, and a reasoned disciplinary decision, except for reprimands and oral warnings; it also permits judicial recourse. Article 99 establishes an officers’ disciplinary council composed of senior Ministry officials. For non-commissioned officers and rank-and-file personnel, Article 118(39) of the implementing regulation expressly identifies improper use of weapons, vehicles, and equipment as a disciplinary offence, while Article 124 establishes disciplinary councils within governorates, services and regular units. These mechanisms operate within the police hierarchy.

External scrutiny is provided through the Public Prosecution and courts. Article 85 of the Criminal Procedure Law places officers exercising criminal-investigation powers under the Prosecutor-General’s supervision and permits requests for examination of breaches of duty and disciplinary proceedings, without excluding criminal prosecution. Article 86 permits referral to the Court of Appeal where misconduct is serious, the disciplinary penalty is insufficient or the administrative authority fails to respond. Following a preliminary investigation under Article 87, the court may issue a warning or temporarily or permanently withdraw the officer’s criminal-investigation powers under Article 88. Under Article 89, nationwide withdrawal entails dismissal, while withdrawal limited to a particular jurisdiction entails transfer out of that jurisdiction.

The NCIAVHR provides a broader external investigative mechanism. Its activities include investigating alleged human-rights violations, inspecting prisons and detention centres and preparing cases for referral to the judiciary. Whether these mechanisms deliver effective accountability for excessive force, including by formations entering the MoI, is examined below.

Police Compliance with the Law of Law Enforcement and Human Rights

The assessment covers the formal police and other government-aligned formations when exercising policing or internal-security powers, including criminal investigation, migration detention and public-order operations. Arrest or detention by any formation does not, by itself, make an operation one of law enforcement: conflict-related captivity and treatment of persons held because of alleged enemy affiliation are assessed under ‘Treatment of persons in the power of the enemy’ above. Mixed military/security caseloads are presented here once for their relevance to detention safeguards, without reclassifying every incident as policing or treating the figures as ordinary-police totals.

Excessive force in policing assemblies

Concerns about unlawful lethal force against demonstrators predate the current conflict. In its report In the Name of Unity, Human Rights Watch documented six occasions during 2008–09 when security forces opened fire on unarmed southern demonstrators, often without warning and at short range, killing at least eleven people and injuring dozens. Its separate investigation, Days of Bloodshed in Aden, examined attacks between 16 and 25 February 2011 that killed at least nine protesters and injured up to 150. Based on more than fifty interviews, hospital records, photographs, videos, and ballistic evidence, the organization documented security forces pursuing and shooting fleeing protesters, obstructing medical assistance, and firing at people attempting to rescue the wounded.

The GEE (A/HRC/42/CRP.1, paras. 607–608 and 613) investigated the killing of demonstrators at al-Husawayn on 13 November 2018. Approximately one hundred unarmed people were protesting against a proposed military encampment. Saudi-backed Yemeni forces deployed a mine-resistant vehicle and up to eleven other military vehicles with mounted weapons. According to the investigation, the protected vehicle entered the gathering and fired its mounted weapon in indiscriminate bursts at close range, killing two demonstrators and seriously injuring another. Firing continued towards protesters who had fled into nearby mountains. Armed local residents arrived and exchanged fire only later. The GEE found reasonable grounds to believe that Yemeni armed forces used excessive force against peaceful demonstrators presenting no apparent threat to life or serious injury, and stated that individual criminal responsibility for war crimes could arise. Although the governor subsequently accepted the need for an independent inquiry, the GEE knew of no steps to establish one.

OHCHR reported unnecessary and disproportionate force by STC-affiliated security forces against demonstrations over deteriorating living conditions and public services. On 15 September 2021, officers used live ammunition after mostly peaceful protests turned violent. Two people, including a child, were killed and several injured when officers opened fire indiscriminately after an unidentified person threw a grenade. OHCHR emphasised that live ammunition could be used only as a last resort against individuals presenting an imminent threat of death or serious injury.

More recent reporting indicates continued restrictions on assembly. Amnesty International’s account of 2025 described excessive force, restrictions on protest movements, and interference with filming during women-led demonstrations in Aden over water and electricity outages. It reported a general ban on demonstrations and public events issued by Aden’s STC security committee on 17 May. The detention of women human-rights defenders during these demonstrations is examined under ‘Arbitrary detention and enforced disappearance’ below.

Human Rights Watch’s March 2026 investigation examined February protests in Aden, Hadramawt, and Shabwa. In Aden on 19 February, verified videos showed protesters breaching a barrier and throwing debris, and security personnel firing upwards, including from an armoured vehicle’s mounted machine gun. Human Rights Watch concluded that government forces used excessive force against peaceful protesters. One death and at least 25 injuries were reported by an STC representative. In Seiyun on 6 February, witnesses described National Shield Forces firing after an unidentified third party opened fire towards protesters and security personnel. It also documented detention at the airport without charges and release conditioned on pledges against ‘unlicensed’ protests.

Unlawful killings outside public-order operations

The NCIAVHR’s Fourth Periodic Report (pp. 76-77) documented the killing of Ahmad Matloob Muhammad Salem al-Harithi during an Aden Security raid on his family’s accommodation in Mualla on 10 December 2017. According to the Commission, personnel shot him, injured his wife, assaulted his daughter and prevented neighbours from providing medical assistance. He was subsequently taken to hospital and pronounced dead. Relatives encountered difficulties lodging a complaint identifying Aden Security personnel. The Commission referred the case to the Ministry of Interior and local authorities, requesting investigation and the arrest of those responsible.

The NCIAVHR’s Twelfth Periodic Report (p. 48) classified the death of Nasser Mohammed al-Huraibi in Maas village, Ataq, Shabwa, on 24 January 2022, as an extrajudicial killing. General and special security personnel demanded that he surrender, opened fire when he refused and left him wounded and bleeding overnight until he died. The Commission attributed responsibility to government general and special security forces and the governorate security department. Its account does not establish whether he was armed or posed an immediate threat.

In its 2019 detailed findings (paras. 262–264 and 276–284), the GEE investigated ten assassinations attributed to the UAE and affiliated Security Belt forces. Eight victims were religious leaders or influential persons associated with al-Islah; witnesses identified Security Belt or UAE personnel in five cases. These included Ramzi Mohammed al-Zughair, a school principal shot repeatedly with AK-47 rifles outside his school in Aden on 23 September 2018, and Ra’fat Danbagh, killed in March 2019 after threats linked to his identification of alleged perpetrators in a rape case. The GEE found reasonable grounds to believe that the killings constituted arbitrary deprivations of life and, where Security Belt personnel acted as law-enforcement officials, extrajudicial executions. It also identified possible criminal responsibility for murder under IHL.

The GEE’s 2020 detailed findings (paras. 100–103 and 109) documented further killings. On 7 June 2019, Security Belt personnel killed five men and injured six at Al-Taweed mosque in Mathad, Al-Dhale’e. Two worshippers were shot inside; three others, including the imam, were taken outside and summarily executed after accusations of Houthi affiliation. Police investigated and sought eight suspects’ arrest, but the GEE reported that they remained unarrested. On 2 February 2020, government Special Security Forces killed two men who approached the requisitioned Al-Aram school in Shabwa to negotiate the release of two detained boys. The GEE concluded that the killings examined constituted arbitrary deprivations of life; some could also constitute war crimes depending on their connection to the conflict.

National investigations continue to identify extrajudicial killings attributed to government forces. The NCIAVHR’s Fourteenth Periodic Report summary, covering August 2025–July 2026, recorded two such incidents. These are incidents investigated during the reporting period, potentially concerning earlier events; the attribution does not establish that both involved ordinary police. The Commission also identified weak enforcement of investigative outcomes and disrupted avenues for remedy. Read alongside the unexecuted arrest requests and resistance to judicial orders documented above, these findings indicate that investigations do not consistently result in suspects being brought before the courts.

Arbitrary detention and enforced disappearance

The Committee against Torture’s 2010 concluding observations (CAT/C/YEM/CO/2/Rev.1, paras. 9–13) identified enforced disappearances, mass arrests without warrants, and prolonged detention without charges or judicial process. It highlighted the multiplicity of security bodies exercising arrest and detention powers without clearly established legal authority.

Amnesty International’s 2018 investigation, “God Only Knows if He’s Alive”, examined 51 men detained between March 2016 and May 2018 in Aden, Lahj, Abyan, Hadramawt, and Shabwa. Most cases involved enforced disappearance; nineteen men remained missing at publication. Amnesty documented detention by UAE forces and UAE-backed Security Belt and Elite formations operating outside the Yemeni government’s effective command. Families searching through police stations, prosecution offices, prisons and coalition bases encountered silence or intimidation, while some detainees were held without charges for up to two years. The investigation therefore concerned both unlawful deprivation of liberty and the concealment of detainees’ fate or whereabouts.

In its 2021 report (A/HRC/48/20, paras. 40–42), the GEE continued to document allegations of arbitrary detention and enforced disappearance involving government and affiliated forces. In Shabwah, it examined allegations concerning six medical workers, including one reportedly seized by armed men in late 2020, transferred to the Criminal Investigation Department and detained for eight days on accusations connected with the treatment of injured patients.

Torture, ill-treatment and sexual violence

In its 2010 concluding observations (paras. 8–10), the Committee against Torture expressed concern about widespread torture and ill-treatment in Yemeni prisons, including facilities operated by security bodies and the MoI’s counterterrorism department. It criticised the infrequency of investigations and prosecutions and the apparent climate of impunity. It also identified inadequate access to lawyers, medical examinations, and judicial review, and insufficient monitoring of security and unofficial detention facilities, conditions increasing detainees’ exposure to abuse.

The GEE’s 2020 detailed findings (para. 168) concerning torture and sexual violence at Marib Political Security Prison and the linked Al-Saleh Institute are detailed under Treatment of persons in the power of the enemy above. facility. The same findings are relevant here to safeguards against torture by security and intelligence personnel.

HRW’s April 2018 investigation documented torture and sexual assault of African migrants, asylum seekers and refugees at the al-Bureiqa detention facility in Aden. Former detainees described guards raping women, girls and boys, beating detainees and withholding food. The facility was formally under government control, and its commander served under the MoI. The Ministry told HRW that it had suspended him and would support investigations, while acknowledging that it did not control the Security Belt units bringing migrants to the facility. The commander denied wrongdoing. This account is particularly relevant because it identifies a formal MoI connection alongside the participation of UAE-backed formations.

The GEE’s 2019 detailed findings (paras. 290–294) found reasonable grounds to believe that Security Belt personnel committed rape and other sexual violence against African migrants and refugees in Aden, including at al-Bureiqa and in Basateen. The findings included gang rape and abduction, the use of weapons to subdue survivors and threats intended to prevent reporting. Its 2020 investigation (paras. 195–198) verified further abuses during March–June 2019: Security Belt personnel raped five women and four girls, subjected fifteen male detainees to forced nudity, and subjected two women and two girls to other sexual violence. These findings concerned ad hoc detention facilities along the Lahj–Aden route and the 22 May Stadium in Aden.

The UN Panel of Experts’ 2021 report (para. 132) investigated sexual violence by Security Belt personnel against two internally displaced women in Dar Sa’d, Aden. It noted that, as of November 2020, no legal procedures had been undertaken despite an initial police report documenting the incident and identifying suspects as security personnel. Its 2022 report (para. 109) also identified evidence that at least ten of fourteen defendants prosecuted in Aden for attacks against mosques and clerics had been tortured and forced to sign confessions; the presiding judge disregarded evidence of that abuse.

Mwatana’s In the Darkness, published in June 2020, attributed 65 torture incidents and fourteen deaths in detention to the IRG, and141 torture incidents and twenty-five deaths in detention to UAE forces and affiliated armed groups during May 2016–April 2020. The report presents torture and deaths in detention as separate categories. These mixed military/security totals complement the facility-specific evidence discussed under Treatment of persons in the power of the enemy above; they should not be added to that evidence as separate cases.

Mwatana’s June 2025 briefing extended its documentation to September 2014–May 2025, attributing 204 torture incidents to IRG forces, 278 to UAE-backed STC forces, forty-two to Saudi/UAE-led coalition forces and six to the Joint Forces on the west coast. It documented beatings, suspension, electric shocks, sexual violence and threats against detainees or their relatives across the wider detention system. One detailed case concerned a nineteen-year-old arrested by STC-affiliated personnel in Al-Dhale’ on 23 January 2023. His father described cable beatings, nail extraction, and electric shocks, including to his genitals, intended to compel confessions. These cumulative figures cover broader actor categories and should not be added to the earlier totals.

The GEE’s final report in 2021 (paras. 72–73, 85 and 87(b)) maintained findings concerning torture and sexual violence attributable, as applicable, to the Yemeni government, coalition actors and the STC. It also reported that law enforcement agencies had refused to investigate and prosecute gender-based violence. Some acts could amount to war crimes where connected to the armed conflict.

Accountability and remedies

 In its 2010 concluding observations (paras. 8 and 16), the Committee against Torture questioned Article 26 of the Criminal Procedure Law, which appeared to require permission from the Prosecutor-General, a delegated public attorney or heads of prosecution before proceedings could be brought against law-enforcement officers or public employees for duty-related offences. It requested clarification of the provision’s continued operation and application, against a background of torture allegations seldom being investigated or prosecuted. The Committee also identified unclear responsibility for investigating complaints and called for investigations independent of the police and military.

A separate obstacle concerns immunity for violations committed under the former regime. The Human Rights Committee’s 2012 concluding observations called for repeal of Amnesty Law No. 1 of 2012, which granted blanket immunity to former President Saleh and immunity for political crimes, excluding terrorism, to those who served under him. Its 2021 list of issues again asked whether the law had been repealed so that perpetrators of serious violations could be held accountable.

In its 2021 report (paras. 74–75), the GEE acknowledged the National Commission’s continuing investigations but reported no increase in criminal prosecutions: nineteen cases reported before the courts in 2020 remained pending. It identified compromised judicial independence, politicisation, inadequate protection of victims and witnesses, and fair-trial deficiencies as obstacles. These figures concern the broader human-rights caseload and should not be presented as police-specific prosecution statistics.

The NCIAVHR’s Fourteenth Periodic Report (pp. 20–26) describes visits to official and unofficial detention facilities, hearings concerning detention, disappearance and torture, and correspondence with government, security, and military bodies concerning investigations and internal accountability. Its official launch summary also identifies weak enforcement of investigative outcomes and disrupted avenues for remedy and reparation, which undermined victims’ confidence and willingness to report violations. That summary urges the government to expedite investigations and prosecutions and judicial and law-enforcement authorities to protect victims, witnesses, and complainants. The summary does not provide police-specific prosecution, conviction or compensation outcomes.

Yemen ratified the 2004 Arab Charter on Human Rights through Act No. 45 of 2008, as confirmed in its 2009 national report to the UN Universal Periodic Review. The Arab Human Rights Committee is empowered to examine state reports, request additional information and issue public observations and recommendations. Its mandate does not include individual complaints.

Taken together, these sources support an assessment of continuing investigative mechanisms but limited demonstrated accountability and access to effective remedies. For formations entering the MoI, institutional incorporation should therefore be assessed alongside compliance with judicial orders, investigation of allegations, sanctions against perpetrators and redress for victims.

Risk of Diversion and Re-Export of Imported Weapons

Targeted arms embargo and scope of the assessment

The UN arms embargo concerns specified recipients, rather than all arms transfers to Yemen. Paragraph 14 of Security Council Resolution 2216 (2015) prohibits direct or indirect supplies of arms and related materiel to or for the benefit of designated individuals and entities, including those acting on their behalf or at their direction, together with associated military assistance, training, and financing. Paragraph 5 and the annex of Resolution 2624 (2022) expressly subjected the Houthis as an entity to these restrictions. Transfers to the internationally recognised government (IRG) are therefore not prohibited simply because Yemen is their destination. AQAP is separately subject to an arms embargo under the UN Al-Qaida sanctions regime; its listing in January 2010 predates the Yemen-specific embargo.

This assessment distinguishes prohibited supplies to embargoed recipients, unauthorised onward transfers by an initial recipient, and subsequent losses through capture, theft or sale. Supply to a government-aligned formation is not automatically prohibited by the Yemen embargo, but may breach the original supplier’s end-user or retransfer conditions. Capture does not by itself prove a violation by the original exporter. These distinctions matter because weapons can enter Yemen through coalition partners or commercial intermediaries and subsequently change hands within or beyond the country.

Coalition onward transfers and loss of end-use control

Earlier UN findings identify inadequate safeguards around coalition supplies. The 2016 Panel report (paras. 84–85 and annexes 15 and 17) reported provision of weapons to armed groups without adequate accountability and secure storage. It collected footage of weapons airdropped into Aden, some subsequently reaching Houthi forces, and documented resistance fighters offering small arms and light weapons for sale through social media.

In its February 2019 investigation, When arms go astray: The deadly new threat of arms diversions to militias in Yemen, Amnesty International documented foreign-made weapons and vehicles held by UAE-backed Yemeni formations, including the Security Belt, Shabwani Elite and Giants. Identified equipment included US-made M-ATV, Caiman and MaxxPro armoured vehicles with mounted heavy machine guns and Serbian Zastava machine guns. Amnesty also identified Belgian Minimi light machine guns used by the Giants, describing these as likely sold to the UAE. The investigation supports concerns about onward distribution and identifies the UAE as a major conduit; it does not establish every intermediate transfer or that each formation possessed every named system.

Diversion through military fragmentation, surrender and capture

Diversion from Yemeni state holdings has occurred through military units’ realignment, seizure of installations, and negotiated surrender. The UN Panel of Experts’ 2016 report (paras. 86–89 and annex 18) documented regular military and security units supplying equipment and assistance to Houthi-Saleh operations. Former Republican Guard and Special Operations Forces/Central Security units systematically aligned with the Houthis. The Panel identified participating brigades and considered named commanders strongly likely to be responsible for their units’ involvement in embargo violations.

The same report documented widespread plundering of State arsenals during 2015, often without resistance. Its Bayhan case concerned the 12 February attack by Ansar al-Sharia and tribal fighters on the 19th Brigade’s base. It also noted reports of seven other brigades being plundered. Officials and activists attributed many incidents to commander-orchestrated handovers, an assessment that should not be treated as verified for every case. Annex 14’s brigade inventory likewise does not establish quantities captured by AQAP (para. 83 and annexes 14 and 18).

The Sana’a Center’s investigation of Mukalla illustrates how negotiated surrender could transfer installations, equipment, and personnel together. Following AQAP’s takeover on 2 April 2015, it seized the Central Security Forces’ camp in Bowaish and its depot, including US-made Humvees, and subsequently the 27th Mechanized Infantry Brigade’s base at al-Riyan. Personnel at Bowaish were offered the choice of surrendering their weapons and leaving, fighting, or remaining to work with AQAP for higher pay; some stayed. AQAP negotiated separately with commanders rather than facing a coordinated military response. The investigation also reports that its April 2016 withdrawal from Mukalla allowed fighters to retain weapons accumulated during their occupation.

Reporting from September 2026 provides a more recent example of battlefield capture. A Reuters investigation published on 17 September described Houthi videos displaying equipment captured from Saudi-backed forces in the Tihama, including armoured and combat vehicles, a tank, artillery, small arms, and a drone-jamming device. This indicates renewed loss of government-aligned equipment, although the reporting does not establish a complete inventory, original export history or precise former custodian of each item.

Insider sales, corruption and illicit markets

The Sana’a Center’s 2018 study of corruption reported sales by senior commanders and soldiers of weapons and military vehicles supplied by the IRG or coalition. Transactions reportedly took place directly, through intermediaries or through local arms markets. Inflated personnel rosters also generated allocations of weapons, ammunition, and equipment exceeding actual staffing requirements, creating opportunities to retain or sell surplus materiel.

The Center’s accompanying analysis, drawing on the same investigation, specifically described anti-Houthi commanders selling surplus weapons across frontlines to Houthi forces. Delayed salary payments have also been associated with weapons sales. A 2018 study published by the Joint Special Operations University cited reports of Yemeni army personnel selling weapons to AQAP or Ansar Allah after months without pay.

TheSana’a Center study described trafficking through eastern governorates nominally controlled by the IRG, facilitated by payments to dealers, transporters, and checkpoint personnel. Transit through government-controlled territory does not establish diversion from government stocks or government authorisation. Nevertheless, commercial networks crossing military and political boundaries provide outlets for diverted equipment and undermine territorial controls.

The 2023 UN Panel report (S/2023/833, paras. 104–106) recorded reports of offers of small arms and light weapons, including man-portable air-defence systems and anti-tank guided missiles, particularly in Houthi-controlled Sana’a and Sa’dah. Type 56-1 and AKS20U rifles offered for sale bore serial numbers consistent with weapons intercepted at sea. The Panel assessed that additional shipments had probably escaped interception.

Diversion through commercial and third-country recipients

The UN 2024 Panel report (S/2024/731, paras. 92–95 and annexes 153–166) traced Brazilian, Turkish, Czech, and Slovak-made pistols advertised in Houthi-controlled areas to earlier recipients in third countries. Annex 156 identifies thirty-seven Taurus G3c pistols transferred to Boomslang Logistics East Africa Limited in Tanzania and reproduces an end-user certificate. Annex 157 identifies twenty Taurus 24/7 pistols sold to Djibouti’s Ministry of Defence, with serial-number lists, sales records, and an end-user certificate. The report also traces weapons through Saudi recipients.

An earlier Panel investigation (S/2017/81, para. 80 and annex 41) examined attempted diversion of 3,000 Taurus pistols and revolvers, part of an authorized sale of 8,000 weapons ostensibly intended for Djibouti’s Ministry of Defence. Saudi authorities intercepted the consignment in 2015, suspecting an onward destination involving the Houthis. The Panel identified the purchaser as Yemen’s Itkan Corporation, owned by Adeeb Mana’a, son of designated arms trafficker Fares Mana’a, and described dubious end-use certification and likely illicit regional-market destinations.

Trafficking out of Yemen and regional spillover

The Global Initiative Against Transnational Organized Crime’s 2022 survey of Somali arms markets documented an apparently Saudi-made G3 rifle in Badhan whose seller said it came through smugglers sourcing weapons in Yemen. Prior supply to anti-Houthi forces was considered probable, but a Saudi response did not establish its export history. The report also examined a July 2021 seizure of fifteen G3 rifles in Puntland, at least four apparently Saudi-made, and a May 2021 seizure of arms and ammunition transported from Yemen. The latter included 30,000 rounds of 7.62 × 51 mm ammunition; photographed packaging indicated Saudi production in 2018. Both consignments were reportedly intended for al-Shabaab. The researchers considered prior supply for the Yemeni conflict highly likely, while noting incomplete tracing responses.

The 2024 Panel report (S/2024/731, para. 88) also identified serial-number correspondences between weapons documented in Somalia and those offered for sale in Houthi-controlled areas or intercepted en route to the Houthis. Such correspondences may indicate a common supplier or transfers in either direction; they do not alone establish the direction of movement. The 2025 Panel report (S/2025/650, paras. 74–81) describes Somali interceptions of explosives and drones moving from Yemen and investigations into trafficking networks. It reports Houthi transfers to al-Shabaab and purchasing cells in Hadramawt. It also identifies Somalia as a transit hub for supplies to the Houthis.

These cases concern illicit onward trafficking rather than demonstrated authorized Yemeni re-exports. They show how Yemen’s conflict and arms markets can contribute to regional supply networks while leaving gaps in the preceding custody chain.

Embargo evasion and the wider trafficking environment

Separate procurement networks continue to supply the Houthis. The UN Panel’s 2025 report, covering August 2024–July 2025, documents continuing smuggling and government seizures. Its para. 107 and annex 79 record government interception on 27 June 2025 of a dhow carrying more than 750 tonnes of munitions and military hardware, including missiles, components, and drone engines. Interceptions also concern components and commercial equipment. The 2024 report (para. 86 and annex 115) records 1,560 Kalashnikov spare parts seized at Aden’s free-zone customs port in November 2023, alongside communications and anti-drone equipment intercepted at border crossings.

National capacity and the integration of armed formations

As discussed, restructuring since January 2026 seeks to integrate formerly separate formations into state structures. The Sana’a Center’s January–March 2026 review reports incorporation of former Security Belt forces into the MoI as National Security Forces, alongside other formerly STC-aligned elements. It also describes a military personnel database and biometric registration linked to salaries, while identifying unresolved loyalties and fragmented authority.

Recent payroll reforms address another vulnerability. An August 2026 report, drawing on officials and soldiers, describes direct salary payments through financial institutions instead of unit commanders. The Defence Minister reported resistance from commanders benefiting from the previous arrangements. These measures are relevant to fictitious personnel and commanders’ control over resources, but do not establish that inflated weapons allocations or insider sales have ended.

Overall assessment

Evidence supports the determination of a high risk of diversion and unauthorized onward transfer, based on documented coalition redistribution, military realignment, seizure and surrender of installations, insider sales, commercial diversion, and trafficking from Yemen to Somalia and Sudan. The pathways concern both Yemeni state holdings and equipment initially exported to coalition or commercial recipients. The targeted embargo is particularly relevant when equipment reaches the Houthis, while AQAP is covered by a separate UN sanctions regime.