The Caribbean Killings

Beginning on 2 September 2025, United States forces deliberately destroyed small vessels in the Caribbean and eastern Pacific and killed the people on board. These were not ordinary Coast Guard interdictions followed by arrest and criminal process. Official US oversight reports state that at least 64 vessels were struck and at least 212 people were killed or presumed dead through 30 June 2026.

This case comes before the study of the January 2026 attack on Venezuela because it came first in the causal sequence. The maritime campaign established a military presence, a vocabulary of exceptional threat, a secret targeting process and an expectation that lethal action near Venezuela would produce criticism but no effective external restraint. The sea was not a secondary theatre after the Venezuelan intervention. It was one of the places where the threshold for that intervention was lowered.

I had already written about that danger in Miedo on 9 December 2025. At that moment the attack on Venezuela had not yet occurred. My fear was not a prediction derived from privileged information; it was an inference from observable conduct. A government that could kill people near my country, publish a label instead of evidence and encounter no material consequence had acquired information about what the international system would tolerate.

In Miedo, I wrote about watching this violence in real time and in high definition. That phrase identifies a variable, not an ornament: visibility. These deaths were announced, justified and then normalised in public. Distance may give a civilian safety; it is not an excuse for pretending that the conduct could not be known.

The people killed were described by the US administration as members of designated drug-trafficking or terrorist organisations. The full evidence, targeting criteria and list of organisations remained classified or unavailable to the public. They were killed before any public charge, trial or independent adjudication of their identities.

The killings did not occur without criticism. They occurred without effective external consequences sufficient to stop them.

That distinction is central to this manual. “Nobody reacted” is empirically false. Officials and institutions did react. The measurable question is whether their response altered the conduct, increased its cost or protected the next potential victim. Here, the publicly observable answer was no.

Method and limits

This study does not infer that every person on every vessel was innocent. It reaches a different conclusion: the public has not been given evidence sufficient to determine individual guilt, and individual guilt would not by itself establish a legal right to kill. Nor does the existence of organised crime prove the existence of an armed conflict. The analysis therefore separates four variables that official rhetoric tends to collapse: suspected criminal activity, organisational designation, the legal classification of hostilities and the necessity of lethal force in each incident.

The official totals are minimums, not a complete casualty registry. “Killed or presumed dead” is retained because that is the language used in the US oversight record. Where identities, nationalities or organisational memberships are not public, this study does not supply them by inference.

What the US government’s own reports establish

The Department of Defense Inspector General’s first quarterly report on Operation Southern Spear states that the first strike, against an alleged Venezuelan cartel vessel on 2 September 2025, killed eleven people. From September 2025 through March 2026, the military struck at least 45 small boats, killing or leaving presumed dead 156 people.

The next report records at least 19 additional vessels and 56 additional people killed or presumed dead from April through June 2026—three vessels in the Caribbean and sixteen in the eastern Pacific. Added together, the official minimum is 64 vessels and 212 dead or presumed dead by 30 June.

The reports also establish what the public does not know. The mission statement was classified. A complete list of designated organisations treated as lawful military targets was not public. Full operational effectiveness data were unavailable publicly. The Inspector General was still evaluating the supporting intelligence and compliance with the Department’s six-step targeting process.

Congress had not enacted a specific authorisation for military force against drug-trafficking organisations. The executive branch instead characterised the campaign through concepts of armed conflict, insurgency, self-defence and transnational threats. This classification was decisive: once the administration treated alleged traffickers as wartime targets, the ordinary model of surveillance, interception, detention, evidence and trial was displaced by lethal attack.

A non-lethal system already existed

The maritime drug trade is not a problem for which governments lacked enforcement tools. The United States had long operated a Coast Guard-led system with international partners to detect vessels, stop them, seize cargo, arrest suspects and present evidence in court.

The contrast appears inside the same official reporting period. From April through June 2026, while military strikes killed or left presumed dead 56 people aboard 19 vessels, the Coast Guard conducted 25 non-lethal interdictions, seized approximately 21 tons of cocaine and detained 45 suspected smugglers.

This does not prove that every military target could have been captured without risk. It does prove that capture and prosecution were functioning alternatives within the same theatre. The government therefore bears the burden of showing why lethal force was necessary in each case, why an imminent threat existed and why interdiction was infeasible.

The international legal objection

International human-rights law protects the right to life and permits intentional lethal force in law-enforcement settings only under strict necessity to protect life. An executive designation does not itself create a non-international armed conflict. That classification depends on facts, including the organisation of the armed group and the intensity of hostilities; trafficking or criminality alone is insufficient.

In October 2025, the UN High Commissioner for Human Rights stated that the strikes violated international human-rights law, said the information publicly available did not show that the people on the boats posed an imminent threat to life, described the deaths as extrajudicial killings and called for the attacks to stop. He called for prompt, independent and transparent investigations and said suspected drug offenders should be arrested, prosecuted and tried.

The European Union’s foreign-policy chief publicly recalled that force under international law requires self-defence or Security Council authorisation. That was a correct statement of the basic inter-state rule. It was not followed by a European response that imposed costs sufficient to halt the campaign.

The designation-to-kill chain

The operational structure can be expressed as five stages:

executive designation: an organisation is placed in a security or terrorist category;
secret intelligence: the government claims a vessel or person falls within that category;
lethal strike: military force destroys the boat;
posthumous label: the dead are described through the designation that justified the attack;
blocked scrutiny: underlying evidence, target list and decision process remain classified, while no defendant survives to contest them.

This is a designation-to-kill chain. It moves the decisive judgment from an independent tribunal to the same executive branch that identifies, attacks and publicly characterises the target.

The word “assassination” is not the most precise term because it often refers to the targeted killing of a prominent individual. The stronger legally grounded description is deliberate state killing without arrest, charge or trial, and—following the UN human-rights chief—extrajudicial killing where no armed-conflict or imminent-threat justification is established.

The Venezuela connection

The first reported target was described as linked to a Venezuelan cartel. The campaign formed part of a broader US military escalation in the Caribbean that preceded the 3 January 2026 intervention in Venezuela. Switzerland’s subsequent legal memorandum stated that the pre-intervention sinking of boats and extrajudicial killing of persons violated international law and occurred before an armed conflict to which international humanitarian law could apply.

Those facts connect the maritime campaign to the coercive environment surrounding Venezuela. They do not prove that every person killed was Venezuelan, worked for the Venezuelan state or even belonged to the organisations asserted by Washington. The absence of public evidence is the point: nationality and guilt cannot responsibly be inferred from an executive label applied after death.

The connection is also chronological. The first strike occurred in September 2025; Miedo was published in December; the United States attacked Venezuela on 3 January 2026; and Dolor, published in April, recorded what it meant to see the earlier fear become an event. The articles are not evidence for the attacks. They document the development of the author’s question: what should a civilian infer when exceptional violence is normalised close to home?

The test civilians and institutions should apply

For every lethal maritime strike, ask:

Who designated the organisation and under what law?
What evidence connected the vessel and each individual to it?
Did a legally cognisable armed conflict exist?
What imminent threat to life made killing necessary?
Could the Coast Guard or a partner have interdicted the vessel?
Were identities verified before the strike?
Were survivors searched for, rescued and treated?
Did an independent body examine the intelligence and proportionality assessment?
Can families obtain the names, evidence and an effective remedy?
Would the United States accept another state using the same secret process against suspected American traffickers?

Until governments answer those questions, “counter-narcotics” is a policy objective, not a legal licence to kill.

The European and Swiss enforcement ceiling

It is inaccurate to say that Europe was silent. European officials stated the applicable rule, and international officials condemned the killings. The relevant question is where their action stopped on the enforcement gradient.

There was criticism, scrutiny and diplomatic language. There were no publicly identified European or Swiss coercive measures directed at the officials, units, assets or cooperation channels responsible for the campaign. The disparity with enforcement against adversaries is therefore again visible: the ally’s conduct is named, but the political relationship is protected.

This is not neutrality. It is selective consequence. The victims receive recognition only after the executive branch that killed them has also defined who they were.

Conclusion

The most important evidence comes from the United States itself. Its oversight reports record the scale of the killing, the absence of a specific congressional authorisation against trafficking organisations, the classified mission and target list, and the simultaneous success of non-lethal interdiction.

The issue is not whether drug trafficking is serious. It is whether an executive can transform suspected criminals into military targets through secret classification and then use death as proof that its classification was correct.

A legal order in which the government names, kills and posthumously judges the same person is not an enforcement system. It is the replacement of adjudication by force.

For me, as a Venezuelan, this was not an abstract defect in maritime law. It was an observable escalation near my country. The next case examines what followed when the machinery previously used against unnamed people on small boats was directed against Venezuelan territory and its government. That sequence—from expendable individuals at sea to coercion of a sovereign state—is why this chapter must come first.