The finding
A collective-defence commitment and an operational defence mechanism are two different things, and the Makkah Joint Defence Agreement currently has the first without publicly having the second.
The commitment is not in doubt. On 7 August 2026 the Saudi Crown Prince and Prime Minister, the President of Türkiye and the Prime Minister of Pakistan signed the Agreement at Al-Safa Palace in Makkah, and the summit joint statement — published by the Saudi Press Agency, by Pakistan’s foreign ministry and through the OIC’s news agency — says that the Agreement is intended to strengthen collective deterrence against any act of aggression, that any armed attack against any one of the three states shall be regarded as an attack against them all, and that it provides for developing all aspects of defence cooperation between them. That is the three governments’ own language about their own instrument. Vigil’s earlier record of this event described the collective-defence clause as secondary-only reporting; that was wrong, and it has been corrected.
Nor is this a summit that produced nothing but a photograph. On 31 August the foreign ministers, the defence ministers and the chiefs of general staff of all three states met in Istanbul as the Agreement’s Strategic Political and Defence Committee. Their joint statement records an intention to institutionalise cooperation further, the establishment of a Secretariat headquartered in Saudi Arabia under a Secretary General, with Pakistan providing the first holder for a three-year term, and commitments on defence-industry cooperation, joint technology development and production, interoperability and cohesion between the three armed forces. Twenty-four days into its existence the Agreement had a standing political-military body and an agreed permanent staff. That is more than most new security arrangements acquire in a year.
What the Agreement does not have in public is the layer between the commitment and a response. The Agreement text has not been released. No implementing protocol has been published. The Istanbul statement names no consultation procedure, no threshold, no command arrangement and no allocation of forces. Türkiye’s foreign minister has said that the degree, form and format of any response would be settled by consultation — which describes an obligation whose content is decided when it arises, not one that follows from a rule.
Eight days after the Istanbul meeting, that gap stopped being theoretical. On 8 September the Houthi movement mounted the heaviest attack on Saudi territory since this round of the war reopened in July. Four days later, no signatory has publicly invoked the Agreement, none has announced a consultation under it, and no act by anyone has been attributed to it by any of the three governments. This note is about what would change that, and why almost everything currently being described as the Agreement’s first test is in fact something else.
What the Makkah Agreement actually established
The Agreement binds three of the largest military establishments in the Muslim world into one defensive formula. Its published undertaking is strong: not consultation, not assistance as each state sees fit, but that an armed attack on one is to be regarded as an attack on all. Read on its own that is Article 5 language, and the comparison is made constantly.
The comparison is also incomplete in a way that matters. NATO’s Article 5 is published, and so is the machinery beneath it: a Council that meets on request, a defined procedure for invocation, a standing command structure and agreed force arrangements. What follows an invocation is still a national decision — the Treaty says each party will take “such action as it deems necessary” — but the route from attack to decision is written down and has been exercised. For the Makkah Agreement none of that route is public. Turkish officials have described a design that resembles it: a ministerial committee, political and military committees, and a permanent secretariat. Describing a design is not the same as publishing a procedure, and the design was described before any of it was built.
Two further characterisations from the parties belong in the record. Turkish officials stated at signature that the pact is not directed at any specific party and is open to other states. Pakistan’s foreign ministry said on 10 September that it is purely defensive, aimed at regional security primarily through deterrence, and that expansion is not currently on the cards because the three states must first consolidate the organisation itself.
Institutions arrived before a tested procedure
The Istanbul meeting is the strongest evidence available in either direction, and it cuts both ways.
It is evidence that the Agreement is being built. Foreign ministers, defence ministers and chiefs of general staff do not meet in the same room by accident, and agreeing a permanent Secretariat with a host state and a named first term is a commitment of budget, staff and prestige. Defence-industrial cooperation and joint technology production are long-horizon programmes that no one begins as a gesture.
It is also evidence of what is missing. Read the Istanbul statement for what it does not contain and the absence is specific: no consultation procedure for an attack on a member, no threshold, no standing command, no force allocation. The Agreement acquired a forum before it acquired a fuse.
Pakistan’s foreign ministry described the condition itself on 10 September, more precisely than any outside analyst could. The spokesperson, Sajjad Haider Khan, called the Makkah Joint Defence Agreement an umbrella agreement, said that under it “we will have to establish subsequent mechanisms to move forward”, observed that when such multilateral organisations are formed “it takes a little while to establish operational details”, and said that significant work lies ahead to finalise the organisation’s structure, mechanisms and procedures. He also said the Agreement is already in force, and that on a military response “there is no such thing under discussion right now”; asked what Pakistan would do, he said “when time comes we will act under the agreement”.
That is the central evidence chain of this note, and it is worth being precise about what it shows. It does not show reluctance. Pakistan’s defence minister, Khawaja Asif, had said the day before that unprovoked aggression against Saudi Arabia, or a spillover from Yemen into the kingdom, “will definitely make this agreement operational”. A government that intended to distance itself from the undertaking would not say that. What the two statements together show is a signatory that accepts the commitment and says, publicly, that the machinery to execute it still has to be established. That is a statement about implementation, not about will — and it is exactly the distinction this note is built on.
Pakistan was already inside Saudi defence
The most common inference in commentary this week has been that Pakistani forces in Saudi Arabia demonstrate that the Makkah Agreement has been activated. The record does not support it, and the reason is a date.
Saudi Arabia and Pakistan signed a bilateral Strategic Mutual Defence Agreement in Riyadh on 17 September 2025, eleven months before Makkah; the governments’ statement said that aggression against either country would be considered aggression against both. On 11 April 2026 the Saudi Ministry of Defense announced the arrival of a military force from Pakistan, including fighter and support aircraft of the Pakistan Air Force, at King Abdulaziz Air Base in the Eastern Sector — and named the instrument: the deployment was made in accordance with the Strategic Mutual Defence Agreement. That is the one foreign military deployment to the kingdom in this war that a government has announced, it happened four months before the Makkah Agreement existed, and it rests on a different treaty. The announcement describes a purpose of joint coordination and operational readiness; it states no combat role, and none is asserted here.
Beneath both instruments sits a defence relationship older than either. Pakistan has trained Saudi personnel for decades and has maintained a military presence in the kingdom across several periods. Pakistan’s foreign ministry said on 10 September that its forces are currently in Saudi Arabia for training and logistical purposes. Figures circulating in reporting for the scale of the present deployment, and for a ceiling in the tens of thousands said to be contemplated under the bilateral pact, are attributed to unnamed officials; this note does not carry them as established and neither does the module.
The analytical consequence is the opposite of the common inference. Because Pakistani forces were already there, on a different legal basis, their presence cannot serve as evidence for the trilateral agreement. It makes attribution harder, not easier — and it means that any claim that Makkah has become operational has to point at something that would not have happened anyway.
One thing is genuinely unresolved here, and the note flags rather than settles it: whether the Makkah Agreement supersedes the 2025 bilateral agreement, incorporates it, or stands alongside it. Different accounts say different things. Pakistan’s foreign ministry said on 10 September that the Makkah Agreement is already in force; it said nothing about the bilateral instrument’s status. Until that is clarified, the April deployment’s attribution stands as the ministry stated it at the time.
Türkiye has shown solidarity, not a Makkah military role
Türkiye’s public position is a single document and it is short. Statement No. 173 of 8 September condemns in the strongest terms the recent attacks carried out by the Houthis against Saudi Arabia, which disregard the safety of civilians and escalate tensions in the region; reaffirms support for Saudi Arabia’s sovereignty and territorial integrity; observes that the attacks undermine diplomatic efforts to end the conflict in Yemen through a lasting political solution; and calls on the Houthis to cease their aggressive actions immediately.
That is political solidarity of a clear and unqualified kind. It is also noticeable for what it omits: the statement does not mention the Makkah Joint Defence Agreement, the instrument under which Türkiye had undertaken a month earlier to regard an attack on Saudi Arabia as an attack on itself. A condemnation issued without reference to a collective-defence agreement, four weeks after signing it, is not a violation of anything — but it is the observable, and it points the same way as the Pakistani briefing.
Turkish diplomacy has been active in the days since: the Pakistani and Turkish foreign ministers spoke on 10 September and reviewed regional developments. No readout of that call describes a Makkah consultation. This note makes no prediction about future Turkish action, and nothing in the evidence supports one in either direction.
Four arrangements, not one
The reason attribution is difficult here is that Saudi Arabia is not waiting for one architecture to mature. It is operating through four at once. Only one of them is reported to have produced anything during the escalation week itself; another had already produced the only announced foreign deployment of this war, five months earlier; a third has a commander and a remit but no announced operation; and the Makkah Agreement produced nothing publicly attributable at all.
The Makkah Joint Defence Agreement is the newest and the one whose operational content is unresolved. The Saudi–Pakistan bilateral agreement is older, has no published machinery, and has already been used. The Multinational Maritime Defense Alliance, declared in Riyadh on 30 July 2026 by fourteen founding states — Saudi Arabia, Türkiye, Pakistan, Egypt, Jordan, Kuwait, Qatar, Bahrain, Sudan, Djibouti, Somalia, the Comoros, Bangladesh and the Yemeni government — exists to keep the Red Sea, Bab el-Mandeb and the Gulf of Aden open to international traffic, and had a named commander, Rear Admiral Abdullah bin Salem Alshehri, appointed on 6 August, the day before the Makkah summit. The United States relationship, resting on the Strategic Defense Agreement of November 2025 and on basing arrangements far older, is the one reported to be supplying targeting intelligence.
Türkiye and Pakistan belong to two of these. That is precisely why a Turkish frigate or a Pakistani aircraft appearing in the Red Sea would not, on its own, be evidence about the Makkah Agreement. The attribution has to be stated by a government, or it is not established — and if every act by a Makkah signatory counts as a Makkah act, the question of whether the Agreement works becomes unfalsifiable.
The Yemen front met Saudi Arabia’s maritime problem
The second half of this story is geographic, and it converged with the first in the space of one week.
This round of the Saudi–Houthi war began on 13 July 2026, when Sanaa International Airport was struck as an Iranian aircraft was expected to return a Houthi delegation from the funeral of the Iranian Supreme Leader. Yemen’s internationally recognised government claimed responsibility; reporting also attributed the strike to Saudi Arabia. The Houthi movement declared its de-escalation phase over and struck Abha airport. On 20 July it declared a blockade on vessels calling at Saudi ports; traffic through Bab el-Mandeb fell by about a quarter in the week that followed, and Saudi Arabia began routing Asia-bound crude through Suez, turning a voyage from Yanbu to South Korea of roughly twenty-four days into one of about fifty-four. From 9 August the Aramco refinery at Jazan was claimed as a target repeatedly.
Then, on 8 September, came the wave. The spokesman of the Saudi-led coalition, Major-General Turki al-Maliki, said attacks on civilian and economic sites at Abha, Khamis Mushait, Jazan and Najran had wounded 73 people including women and children, and that fires had broken out at oil facilities and utilities in the southern region, where operations were temporarily suspended. The Houthi movement said it had struck Aramco facilities and King Khalid air base with dozens of ballistic missiles and drones, in answer to more than a hundred Saudi air strikes in Yemen over three days.
Two days later the ground moved. The Houthi movement took the port of Mokha — about 75 kilometres north of Bab el-Mandeb — at dawn on 10 September and, over the following day, Dhubab some 46 kilometres further south, the coastal villages beyond it, Perim island in the strait and Zuqar in the Hanish group. That is the last stretch of Yemen’s Red Sea coast the internationally recognised government held; several outlets, and a Yemeni security official quoted by one of them, described the result as the whole of that coastline passing into Houthi hands, and no independently verified control line has been published. The movement claims six districts and 5,400 square kilometres. Reporting places a remaining government force on the Hanish islands.
The Yemeni government’s side of this needs four things kept apart, because coverage has run them together. There is a STATED INTENTION: field commanders in Taiz and on the western coast have said they mean to return to Mokha and retake it, and the government has said it will fight the movement until it is eradicated. There is an OPERATIONAL PREPARATION: the Presidential Leadership Council’s vice chairman, Tareq Saleh, whose National Resistance Forces held the coast, said he had ordered a strategic retreat and the formation of an alternative command centre in a safe location to regroup the forces, in consultation with the Arab coalition command. There is no reported COUNTEROFFENSIVE. And there is no TERRITORIAL REVERSAL: nothing taken since 10 September has been reported recovered. An intention is not a preparation, a preparation is not an offensive, and an offensive is not ground regained. This note records the first two and states the absence of the second two.
What that does and does not mean requires care, because the coverage has not taken it. Bab el-Mandeb has two shores; the other belongs to Djibouti and Eritrea, and the wider of the strait’s two channels lies west of Perim. A Houthi politburo member stated that freedom of navigation and international trade in the Red Sea and Bab el-Mandeb are safe and orderly, and that the exception remains Saudi shipping — a belligerent’s declaration, recorded as such. A Houthi commander said that with the fall of Mokha the strait could be considered under the movement’s administration, which is a claim of the same kind. Al Jazeera reported no official Houthi announcement of control over the strait. Andreas Krieg of King’s College London, quoted on the day, put the military reading plainly: the western littoral remains outside Yemen, so the positions do not confer full control, though they do confer a stronger ability to contest and threaten the waterway — and he judged the movement unlikely to sustain a blockade of it.
Shipping has nonetheless moved. One count put transits through the strait at twelve on the evening of 10 September against twenty-nine or thirty on a normal day, and traffic had already fallen by about a quarter after the July blockade declaration. That is a change in what shipowners and underwriters choose to do, and it is the thing that actually constrains a route; it is not evidence of an act against a vessel. No independently observed interference with non-Saudi shipping from the new positions has been read. What has changed is physical access to the Yemeni side of the strait, and the risk that access implies. What has not been established is interdiction.
Access matters even so, because of what the strait is for. Saudi Arabia’s east–west pipeline exists to move crude away from the Strait of Hormuz and deliver it to the Red Sea; between March and mid-July 2026 Saudi seaborne exports through Bab el-Mandeb ran at eight times the level of the same period a year earlier. The bypass had become the route.
And then, on the same morning Mokha fell, the bypass itself was hit. The pipeline was struck in the Riyadh and Medina areas on Thursday 10 September. The Ministry of Energy said the line had been shut down as a precautionary measure after multiple attacks, and that there were injuries requiring medical care; no restart has been announced. The foreign ministry said the following day that the drones had originated from Iraq, that Saudi Arabia would not respond at this stage at the request of the Iraqi prime minister, and that it reserved the right to take all necessary measures to protect its sovereignty, security and critical facilities. Baghdad said its investigations showed the drones were launched from Maysan province, dismissed the commander of the Maysan Operations Command and ordered an investigative council. No group claimed the attack. Satellite-based analysis published the same week reported substantial impacts at two pumping stations and extensive fire damage at one; no operator assessment of the damage has been published, and this note does not state how long the line will be down. Crude traded above $100 through the week.
The analytical point is not that one actor now controls both chokepoints. That claim would outrun the evidence badly in both directions — Hormuz and Bab el-Mandeb are contested by different parties on different terms, and neither is closed. The point is narrower and better supported: pressure is accumulating against every route Saudi Arabia uses to reduce its dependence on Hormuz, and within forty-eight hours — the same Thursday morning, and the Friday after it — it arrived at the middle of that route and at its mouth.
Iran’s role requires careful attribution
The temptation to write “Iran closed the second strait” is strong and should be resisted.
What this module holds as established is a long-reported material-support relationship between Iran and the Houthi movement, whose current content it cannot verify and whose degree of direction it does not assert. That record was written before this week precisely because the movement’s 2026 conduct is evidence against reading the relationship as control.
What was added this week is reported, not established. United States officials are reported to believe that hundreds of Islamic Revolutionary Guard Corps officers are in Yemen, helping coordinate operations aimed at the strait. Five military sources from the internationally recognised Yemeni government told reporters that the IRGC had directed the Red Sea coastal campaign, citing intercepted communications and captured prisoners, and named a senior commander, Abdolreza Shahlaei. Those sources are a belligerent government’s; the assessment is an allied government’s; neither is independent verification.
Iran’s position is a denial. Its foreign ministry spokesman, Esmail Baqaei, rejected the characterisation of the movement as an Iranian proxy on 10 September, saying that Ansar Allah is an independent Yemeni actor that takes orders from no one.
Vigil’s grading is therefore unchanged by the week’s reporting: support is reported, direction is unestablished, control is not asserted. “Iran-aligned” and “Iran-commanded” are different claims and only the first is carried here.
Why Saudi reliance on Washington does not answer the Makkah question
Saudi Arabia asked the United States for direct military help against the Houthi movement and did not get it. The Crown Prince was reported to have raised the question with the American President twice on 10 September. The administration’s answer, according to reporting the same day that rests on unnamed United States officials rather than on any published statement, was expanded intelligence and geospatial targeting support for Saudi strikes, but no direct American strikes, and no refuelling of Saudi aircraft. The CENTCOM commander travelled to the kingdom that week.
It is tempting to read this as a verdict on Makkah — either that Saudi Arabia turned to Washington because the Agreement is empty, or that the American refusal is what makes the Agreement necessary. Neither inference survives contact with the evidence. Saudi Arabia has run its defence through the United States for decades and through a bilateral Pakistani agreement since 2025; using both during an attack is what a state with overlapping arrangements does, and Saudi officials said at the Agreement’s signature that it does not come at the expense of ties with other allies. What the American episode shows is that the security architecture around the kingdom is layered. It shows nothing, in either direction, about whether the Makkah Agreement has a mechanism — and the evidence for that has to come from the Agreement’s own signatories.
What would settle it
The question this note leaves open is not “will they fight”. It is narrower, testable, and belongs in the module’s standing coverage rather than in an article: what would demonstrate that the Makkah Agreement has moved from a collective-defence commitment to an operational defence mechanism.
The answers are institutional and attributional, not military. A signatory formally invoking the Agreement after an attack. Publication of an implementing protocol, or of the Agreement text. The Saudi-based Secretariat standing up, with a seat, a staff and a named Secretary General. A joint command or coordination cell stated to exist under this Agreement rather than under the maritime coalition. A Turkish or Pakistani contribution that a government says is made under the Agreement. A signatory defining what constitutes an armed attack for its purposes.
None of those has happened. Several of them could happen without any shot being fired, and one of them — publication of the text — would settle more than a year of operations. Until one does, the honest description of the Makkah Joint Defence Agreement is the one its own signatory gave on 10 September: an umbrella, under which the mechanisms still have to be built.