When unknown gunmen on a motorcycle shot a Lashkar-e-Taiba commander outside a mosque in Punjab and vanished into traffic before the body hit the pavement, much of the commentary treated the killing as something new. It was described as a sudden departure, a doctrine invented overnight, a campaign that materialized from nowhere. That framing is wrong, and the error matters. The men who plan operations of this kind do not begin from a blank page. They begin from files, from networks of informers built over decades, from courier routes already mapped, from safe addresses already known. The lethal effort that has eliminated dozens of wanted men on Pakistani soil since 2021 did not start in 2021. It began fifty years earlier, in a war that proved India could reach inside its rival and reshape the map of South Asia.

This is the story of how the Research and Analysis Wing came to operate inside the one country it was, in a sense, created to watch. The agency’s footprint in Pakistan is the longest-running and least understood chapter of its institutional life. Pakistani officials describe a vast Indian penetration, agents in every city, sleeper cells in every province, a hidden hand behind every misfortune. New Delhi describes nothing at all, because it has never confirmed a single mission. The truth, as it almost always does in the trade of espionage, sits in the contested space between a government that exaggerates and a government that denies. What follows is an attempt to map that space honestly, era by era, from the founding triumph of 1971 to the shadow war of the present decade, and to argue a single proposition. The lethal effort of today rests on an inheritance, and the inheritance is the real subject.

Few subjects in South Asian security are as distorted by motivated reasoning as this one. Read the Pakistani press and the reader encounters an India that is omnipotent and omnipresent, the unseen author of every separatist bombing, every sectarian massacre, every economic crisis. Read the Indian press and the reader encounters a country that has never lifted a finger across the border, a blameless democracy slandered by a failing state. Neither portrait survives contact with the historical record. The honest account is harder to write and less satisfying to read, because it requires holding two ideas at once, that India almost certainly does operate inside Pakistan, and that the scale of those operations is almost certainly far smaller than Islamabad claims. This article attempts that harder account.

Background and Triggers

To understand why India’s external intelligence service operates inside Pakistan, begin with why the service exists at all. The Research and Analysis Wing was carved out of the Intelligence Bureau in 1968, a year after a humiliating intelligence failure had contributed to military reversals against China. Rameshwar Nath Kao, the founding chief, built the organization on a simple recognition. India needed an agency that looked outward, that gathered and acted upon information beyond the country’s borders, that was not distracted by the domestic policing function the Intelligence Bureau had always carried. The model was external, professional, and quietly ambitious. Kao admired the tradecraft of the Western services and the discipline of Israel’s Mossad, and he wanted an Indian instrument that could match them within a generation. The full institutional account of that founding, the early recruitment, and the slow construction of a global station network is told in the agency’s complete history, and the formal evolution of the rules under which it has operated is examined in India’s covert operations doctrine.

Pakistan was the obvious priority from the first day. The two states had fought wars in 1947 and 1965. They contested Kashmir. They shared a long, porous, militarized frontier. Any external intelligence organization worth its budget would have made the western neighbor the center of its attention, and Kao’s did. Yet there is an important distinction between watching a country and operating inside it. Watching means collection, the patient accumulation of facts about an adversary’s military dispositions, political fractures, and economic stresses. Operating inside means something more, the placement of human assets, the running of agents, the conduct of activity designed to produce an effect rather than merely a report. The history that follows is the history of how the agency moved, unevenly and across decades, from the first kind of work to the second.

Triggers for that movement were rarely chosen freely. India’s intelligence posture toward its neighbor has almost always been reactive, shaped by events the country did not control. The collapse of East Pakistan in 1971 created the first great opportunity. The rise of the Pakistani-backed insurgency in Punjab and then in Kashmir created the pressure that pushed the agency toward harder methods. The Mumbai attacks of 2008 exposed how little defensive intelligence alone could accomplish. Each of these moments reset what the organization was willing to attempt, and each left behind infrastructure, contacts, and lessons that the next phase inherited. New Delhi did not draw up a fifty-year plan to arrive at the motorcycle killings of the present. It arrived there through a long sequence of responses, and the path can be reconstructed if one is willing to follow it carefully.

A second point belongs in this background. India’s covert posture has always been constrained by the country’s self-image. New Delhi has wanted, throughout, to be seen as a status quo power, a responsible democracy, a state that does not behave the way its rival behaves. That self-image has functioned as a brake. It explains why the agency’s offensive capability lay dormant for long stretches, why political leaders repeatedly pulled back from authorizing harder action, and why the present campaign, whatever its true scale, is conducted with such absolute deniability that the government will not even acknowledge it exists. The record of Indian operations inside Pakistan is therefore not a smooth curve of escalating aggression. It is a stop-and-start story, a record of capability built, set aside, rebuilt, and only recently, if the allegations hold, fully unleashed. The discussion that follows traces five distinct eras, and the boundaries between them are the moments when the brake was released or reapplied.

Worth stating here, before the chronology begins, is the central interpretive claim of this article. The shadow war is treated, almost universally, as a rupture. Commentators describe a doctrine that appeared from nowhere around 2021, an aggressive new India shedding its old caution. That reading is comfortable, because it is simple, and it is wrong. A campaign of targeted killing depends on three things. It depends on knowing exactly who to kill and where to find them. It depends on a political leadership willing to authorize lethal action abroad. And it depends on a method of execution that protects the sponsoring state from exposure. Each of these three foundations was laid in a different earlier era. The targeting knowledge was built in the defensive decades. The political will was forged in the fury after a single catastrophic night in 2008. The deniable method, the use of local hands rather than one’s own officers, descends in an unbroken line from the very first major operation, in 1971. Understanding the shadow war means understanding its three inheritances, and the only way to see the inheritances is to walk through the eras one at a time.

The 1971 War and the First Penetration

The first era of Indian intelligence activity inside the territory that is now Pakistan, and the territory that was then East Pakistan, runs through the year that produced Bangladesh. It is the foundation of everything that came after, and it established the template the organization would measure itself against for the next half century.

When the Pakistani military launched its crackdown in East Pakistan in March 1971, it set in motion a humanitarian catastrophe and a strategic opportunity in the same act. Millions of refugees poured across the border into India. A Bengali resistance coalesced, ragged at first, furious, and desperate for support. Kao’s young agency recognized immediately what New Delhi had been handed. Here was a chance to dismember the rival state, to convert its eastern wing into an independent and friendly nation, and to do so with a domestic insurgency that already had every reason to fight. The agency’s task was to organize, train, arm, and direct that insurgency, and to feed the Indian military the intelligence it would need when the conventional war finally came.

The Mukti Bahini, the Bengali liberation force, became the instrument. Indian intelligence officers helped structure its command, ran training camps on Indian soil, supplied weapons and communications equipment, and embedded the resistance within a coherent strategy rather than letting it burn itself out in uncoordinated raids. The agency also ran agents deep into East Pakistan, mapping the disposition of Pakistani garrisons, identifying which commanders were demoralized and which would fight, and tracking the movement of reinforcements. When the Indian Army crossed the border in December 1971, it did not advance blind. It advanced into a theater that had been prepared. The conventional campaign that followed was swift, and the eastern wing of Pakistan surrendered in under a fortnight. A new nation was born, and the rival state lost more than half its population and a vast portion of its strategic depth. The detailed reconstruction of that operation, phase by phase, is the subject of RAW’s first great success.

Dwelling on what the 1971 experience taught the organization is worthwhile, because the lessons traveled forward. The first lesson was that India could reach inside Pakistan and produce a strategic effect of the highest order, not merely a report but a redrawn map. The second lesson was about method. The agency had not sent its own officers to fight. It had identified a local population with its own grievances, supplied that population, trained it, and directed it, while keeping Indian hands at one remove. This indirect model, the cultivation of a proxy rather than the deployment of one’s own shooters, would dominate Indian thinking for the next two decades. The third lesson concerned intelligence preparation. The military victory had been possible because the ground had been mapped, the enemy’s order of battle understood, the weak points identified. Preparation, patient and unglamorous, was the multiplier that turned a war into a rout.

There is a fourth lesson, and it is the uncomfortable one. The 1971 success was, in significant part, a gift of circumstance. Pakistan’s own military brutality had created the refugees, the resistance, and the international sympathy. The Bengali population had its own deep and authentic desire for independence, a desire rooted in years of political and economic discrimination by the western wing. The agency was skilled, and its contribution was real, but it had not manufactured the conditions for victory. It had exploited conditions that history handed it. Honest officers within the organization understood this even in the moment of triumph, and the understanding bred a certain caution. A temptation existed after 1971 to believe that the model could be reproduced anywhere, that any restive minority inside Pakistan could become another Mukti Bahini. The next era would test that belief, and it would not pass the test cleanly.

A quieter consequence of the 1971 campaign was organizational. A young service that had existed for only three years suddenly found itself credited with a role in one of the decisive military outcomes of the postwar era. That credit brought prestige, brought budget, and brought political access, and it brought something more dangerous, a reputation to protect. An institution that has tasted a defining success acquires an appetite for repeating it, and it also acquires a fear of the failure that would tarnish the legend. Both impulses, the appetite and the fear, can be traced through the decades that followed. Appetite pushed planners toward bold proposals. Fear pulled political leaders toward caution whenever a bold proposal carried a real risk of exposure. The tension between those two forces is, in a sense, the engine of this entire history, and it was wound up in the triumph of a single year. A reader who keeps that tension in mind will find the stop-and-start rhythm of later decades far less puzzling, because the rhythm is simply the two forces taking turns.

Consider, too, the international dimension of the 1971 operation, because it shaped how India would think about deniability ever after. The United States, in that year, tilted sharply toward Pakistan, dispatching a naval task force toward the Bay of Bengal as a signal of displeasure with India. The Soviet Union balanced that pressure through a treaty with New Delhi. The point for this history is that India learned, in 1971, that operating against Pakistan carried real costs in great-power relations, that the world watched, and that even a justified intervention invited condemnation from powerful capitals. This early exposure to the diplomatic price of action helped instill, in the Indian system, the instinct that would later define the shadow war, the instinct toward absolute concealment. A state that had been lectured by Washington in 1971 would, decades later, prefer to conduct its hardest operations in a silence so complete that no foreign ministry could be forced to react.

For the purposes of this history, the significance of 1971 is that it established the agency’s appetite for operating inside Pakistani territory and proved that such operations could succeed spectacularly. It also created, in the western wing of the diminished Pakistani state, a permanent and bitter enemy. Islamabad emerged from 1971 convinced that India was an existential threat that would, given any opening, complete the dismemberment. That conviction shaped Pakistani strategy for decades, drove the long sponsorship of militancy as a counter, and ensured that the intelligence contest between the two states would never cool. The first era, in other words, did not only show India what it could do inside Pakistan. It guaranteed that India would always have a reason to keep doing it, and it guaranteed that Pakistan would always be watching for the next move.

The Sindh and Balochistan Question of the 1980s

The second era is the most contested, the most denied, and the most consequential for the way Pakistan understands the Indian threat. It runs roughly through the 1980s and into the early 1990s, and it concerns the allegation that India responded to Pakistan’s sponsorship of insurgency on Indian soil by sponsoring insurgency on Pakistani soil in return.

Background here is essential. Through the 1980s, Pakistan’s military intelligence establishment became deeply involved in two projects that alarmed New Delhi. The first was the Afghan jihad, where the army’s intelligence service managed the flow of weapons and fighters against the Soviet occupation, an enterprise that built an enormous covert infrastructure and a generation of trained militants. The second, and the one that struck India directly, was the cultivation of separatist militancy in the Indian state of Punjab. The Khalistan movement, which sought an independent Sikh homeland, received sanctuary, training, and arms from across the border. Indian Punjab burned through much of the decade. New Delhi watched a neighboring intelligence service run a proxy war on Indian territory and concluded that the contest could not be fought on defense alone.

What India allegedly did in response is the heart of the controversy. Pakistani accounts, and a number of Indian accounts written by retired officials, describe a decision taken at the highest levels in the mid-1980s to answer Pakistan’s Punjab operation with a mirror operation inside Pakistan. The two areas of focus, according to these accounts, were the southern province of Sindh, where the urban politics of Karachi and the rural grievances of the interior offered openings, and the western province of Balochistan, where a long-running Baloch nationalist movement had every reason to resent the central state. The claim is that Indian intelligence cultivated contacts within these movements, provided a measure of support, and used the resulting pressure as leverage, a way of signaling to Islamabad that the proxy game could be played in both directions.

Here the historian must be careful, and the careful reader should be skeptical of every confident assertion. Evidence for large-scale Indian sponsorship of insurgency inside 1980s Pakistan is thin, fragmentary, and heavily dependent on interested sources. Pakistani officials have every incentive to magnify the Indian role, because doing so converts genuine domestic grievances into a foreign conspiracy and absolves the central state of responsibility. The Baloch movement, in particular, has authentic and longstanding causes rooted in the province’s poverty, its political marginalization, the appropriation of its natural resources, and the heavy hand of the security forces. To attribute that movement to Indian manipulation is to insult the Baloch and to misread the conflict entirely. At the same time, the Indian accounts that describe a Sindh and Balochistan operation come from former officials who were positioned to know something, and the strategic logic of a tit-for-tat response to the Punjab proxy war is entirely coherent.

An honest conclusion is therefore a hedged one. India probably did establish contacts and probably did provide some support in this period, the support was likely modest rather than decisive, and the operation functioned more as leverage and signaling than as a serious attempt to dismember the Pakistani state. Note the asymmetry between this conclusion and the two propaganda narratives. Pakistan claims India built and ran the Baloch insurgency. India claims it did nothing whatsoever. The defensible middle is that India touched the edges of movements that already existed, for reasons that already existed, in order to raise the cost of Pakistan’s own behavior in Punjab. That is a meaningful claim, but it is a far smaller claim than either capital would prefer to make.

Worth dwelling on, too, is the urban dimension that the Balochistan focus tends to obscure. Karachi in the 1980s was a city of extraordinary political violence, its neighborhoods fractured along ethnic and linguistic lines, its streets contested by the armed wings of rival movements. A foreign service looking for leverage did not need to manufacture instability in such a place. It needed only to identify which existing fracture might be widened, which aggrieved faction might welcome a quiet friend. Whether Indian intelligence actually did so in Karachi is, like everything in this era, unproven. But the point worth retaining is structural. Covert leverage of this kind rarely creates a conflict from nothing. It locates a conflict that already exists, for reasons rooted in genuine local injustice, and it adds a measured weight to one side of the scale. That is a smaller and more cynical activity than the grand sponsorship that propaganda describes, and it is also far harder to detect, because the foreign hand never has to do very much. A patron who supplies a movement with its weapons and its strategy leaves fingerprints everywhere. A patron who merely encourages, advises, and occasionally funds leaves almost none, and the difference between those two roles is the difference between the campaign Pakistan alleges and the campaign that probably occurred.

Less contested is the fact that the leverage worked, at least in part, and that it worked through negotiation rather than victory. By the late 1980s and into the early 1990s, the two states arrived, through quiet diplomacy, at a partial and unstable understanding. India would wind down whatever it was doing in Sindh and Balochistan, and Pakistan would, in theory, reduce its support for the Khalistan militancy. The Indian side of that bargain appears to have been honored more fully than the Pakistani side, which is a pattern the relationship would repeat across the decades. The Khalistan insurgency was eventually defeated, but the defeat owed more to a brutal and effective police campaign within Indian Punjab than to any reduction in Pakistani sponsorship.

A further dimension of the 1980s deserves attention, because it complicates the simple story of tit-for-tat. India in this period was also deeply entangled in Sri Lanka, where it trained and armed Tamil militant groups in a proxy operation that would end catastrophically. The painful arc of that involvement, which culminated in Indian soldiers dying to fight the very group India had armed, is examined in RAW’s most controversial operation. The Sri Lankan disaster overlapped in time with the alleged Pakistan operations, and the two experiences taught the same lesson from different directions. Proxy sponsorship was a tool that turned in the hand of the state that wielded it. Backing an insurgency created an actor with its own agenda, its own grievances, and its own capacity to outlast the patron’s strategic purpose. Indian planners who lived through both the Punjab proxy war they suffered and the Sri Lankan proxy war they inflicted emerged with a deep and lasting suspicion of the entire model.

The second era left two inheritances that matter for this history. The first is infrastructure. Whatever contacts the agency built in Sindh and Balochistan in the 1980s, whatever understanding it developed of those provinces and their fractures, did not evaporate when the operation wound down. Networks of this kind degrade slowly. Some assets are lost, but relationships persist, knowledge persists, and the map of who can be reached and how is retained in files. The second inheritance is doctrinal. The decade taught Indian planners that proxy sponsorship inside Pakistan was possible but also dangerous and morally fraught, that it invited a symmetrical accusation, and that it produced leverage rather than decisive results. That lesson, combined with the harder lesson of Sri Lanka, would steer the agency away from the proxy model and toward something more surgical. The relationship between these alleged Balochistan operations and the broader institutional record is examined further in the agency’s extended history.

The Counter-Intelligence Decades of the 1990s and 2000s

The third era is, on the surface, the quietest, and that quietness is itself the point. Through the 1990s and the first decade of the new century, Indian intelligence activity toward Pakistan was dominated not by dramatic operations but by the grinding, defensive, and largely invisible work of collection and counter-intelligence. To call this era quiet is not to call it empty. It was the period in which the organization absorbed a series of hard lessons, suffered real failures, and built much of the human and technical apparatus that the later offensive era would draw upon.

Two developments shaped the period. The first was the eruption of insurgency in Kashmir at the end of the 1980s, which transformed the intelligence contest. Pakistan’s military establishment, freed by the Soviet withdrawal from Afghanistan and flush with trained militants and covert infrastructure, redirected its proxy machinery toward the Kashmir Valley. Through the 1990s, organizations sheltered and supported from across the border, including the rising Lashkar-e-Taiba and later Jaish-e-Mohammed, sent fighters into Indian Kashmir in a sustained campaign. The Indian intelligence task became overwhelmingly defensive, the interception of infiltration, the penetration of militant organizations, the disruption of plots before they matured. Officers of the agency spent the decade running sources inside the militant pipeline, and the pipeline ran directly back to Pakistani soil. The structure of that pipeline, its training camps and financing and recruitment, is mapped in Pakistan’s terror safe haven network.

A second development was the maturing of the intelligence relationship into a recognizable spy-versus-spy contest. Both states ran agents against the other. Both suffered penetrations and exposures. The 1990s and 2000s produced a steady churn of arrests, expulsions of diplomats accused of espionage, and the occasional public scandal. This is the texture of the era, and it deserves emphasis because it is so often overlooked. The dramatic histories of intelligence focus on operations that produce explosions or headlines. The reality of the trade is mostly the patient, undramatic management of human sources, the cultivation of a clerk who can photograph a document, a soldier who will describe a deployment, a businessman who travels and observes. The agency did this work inside Pakistan throughout the period, and it did so with mixed results.

Consider the rhythm of the diplomatic expulsions that punctuated these years, because that rhythm reveals how the contest actually functioned. Periodically, one capital would declare a member of the other’s diplomatic mission unwelcome, accusing the official of activities incompatible with diplomatic status, which is the conventional language for espionage. The other capital would respond within days, expelling an officer of equivalent rank. These reciprocal expulsions were rarely about a single exposed operation. They were a form of communication, a way for each intelligence establishment to signal that it had detected the other’s activity, to impose a modest and survivable cost, and to manage the contest within understood limits. The expulsions almost never escalated into a true rupture, because both sides understood the game and both sides valued its predictability. That predictability is itself a finding worth pausing on. For all the genuine hostility between the two states, their intelligence services spent these decades conducting a contest with informal rules, and the rules largely held. What makes the later shadow war so significant is precisely that it broke this older understanding. A reciprocal expulsion is a move within an agreed game. A reciprocal killing is the abandonment of the game itself, and the third era is valuable in this history because it shows, by contrast, just how far the contest would later travel.

Failures of the era were real and instructive. Indian intelligence did not see the Kargil intrusion of 1999 coming, when Pakistani soldiers and militants occupied heights on the Indian side of the line in Kashmir. The lapse was a collection failure, a case of an adversary’s preparation going undetected, and it forced a serious internal reckoning. The agency also struggled, throughout the period, with the central problem of running human sources inside a hostile and heavily policed state. Pakistan’s own counter-intelligence apparatus was formidable, and an Indian asset discovered inside Pakistan faced interrogation, imprisonment, or worse. The price of the trade was paid by individuals whose names rarely became public, and the casualty rate among human sources in a hostile country is a cost that the dramatic narratives almost never count.

Yet the era was also one of quiet construction, and the construction is the part that matters most for everything that followed. Technical collection improved through the period, the agency’s ability to intercept communications and to read the electronic signature of the adversary’s military expanding with each year. Human networks, though battered, persisted and in places deepened. Crucially, the organization spent these decades building an intimate understanding of the very militant ecosystem that the later campaign would target. To penetrate Lashkar-e-Taiba defensively, to track its commanders, to map its safe houses and its couriers and its financing, the agency had to learn the organization from the inside. Such knowledge was acquired for defensive reasons, to stop attacks on Indian soil. But knowledge, once acquired, can be turned to other uses. The files compiled in the 1990s and 2000s to defend Kashmir would, in the next decade, become the targeting material for a different kind of campaign. The bilateral contest of these years, the long duel of penetration and counter-penetration, is examined in detail in the intelligence war between the two agencies.

Pause on that transformation, because it is the hinge of the entire history. A file on a militant commander, compiled to predict and prevent his next attack, contains exactly the same information that would be required to kill him. It records where he lives, which mosque he attends, what vehicle he drives, who his bodyguards are, what his daily routine looks like, which couriers carry his messages. Defensive intelligence and offensive targeting draw on an identical body of knowledge. The only difference between the two is the decision about what to do with the file. Throughout the third era, India compiled the files and used them defensively, because that was the policy and that was the limit of the authorization. The capability to use them otherwise was being created, year after year, as an unintended byproduct of the defensive mission. Nobody in the 1990s decided to build a kill list. The kill list assembled itself, page by page, inside the defensive archive.

The third era closed without a single dramatic Indian operation inside Pakistan, and that absence was a policy choice rather than a failure of capability. New Delhi, through the 1990s and 2000s, repeatedly declined to authorize offensive action. Political leaders of different parties arrived at the same conclusion, that the risks of an exposed operation, the diplomatic cost, the danger of escalation between two nuclear-armed states, outweighed whatever satisfaction an offensive strike might bring. The agency had assets, knowledge, and growing capability. What it did not have was authorization. The brake described in the background section was firmly applied throughout these years. It would take a catastrophe to release it.

The Post-26/11 Turn Toward Offensive Intelligence

A new era opens on a single night in November 2008, when ten gunmen trained and dispatched from Pakistani soil came ashore in Mumbai and conducted a three-day massacre that killed more than one hundred and sixty people. The Mumbai attacks did not change Indian intelligence capability overnight. What they changed was the conversation about what that capability should be used for, and that conversation, once started, did not stop.

The attacks were a profound humiliation for India’s security establishment, and the humiliation operated at several levels. There had been intelligence, fragments of warning, indications that a maritime attack on Mumbai was being prepared, and the system had failed to assemble those fragments into a usable picture or to act on the picture it had. Defensive intelligence, the entire premise of the third era, had been tested at the highest stakes and had not held. Worse, in the aftermath, India watched the perpetrators escape every consequence. The surviving attacker was tried and eventually executed, but the planners, the trainers, the handlers, the men who had conceived and directed the operation from Pakistani soil, faced nothing. The legal process inside Pakistan against the Lashkar-e-Taiba leadership stalled, then collapsed, then became a transparent farce. India had presented evidence, had made its case to the world, had relied on the machinery of international pressure and bilateral diplomacy, and had watched that machinery deliver precisely nothing.

A lesson that a section of the Indian strategic community drew from this experience was stark. Reactive engagement had failed. Presenting evidence had failed. Waiting for Pakistan to act against its own proxies had failed, because the proxies were not a deviation from Pakistani strategy but an instrument of it. If the men who planned mass-casualty attacks on Indian cities were going to face consequences, the consequences would have to be delivered by India, and delivered in a way that did not depend on Pakistani cooperation, international courts, or diplomatic goodwill. This was the intellectual origin of the offensive turn, and it is essential to understand that it was a turn in policy thinking before it was a turn in operations. The doctrinal evolution from a defensive posture toward an offensive one, and the specific failures that drove each step, are traced fully in India’s covert operations doctrine.

Years immediately after Mumbai were spent rebuilding. India overhauled its intelligence architecture, created new institutions for coastal security and counter-terrorism coordination, and pushed for better fusion of information across agencies. The external intelligence service, for its part, was tasked with developing the capability to act, even if the authorization to act did not yet exist. Through roughly 2008 to 2014, the organization deepened its networks, sharpened its understanding of the militant leadership it might one day need to reach, and began to develop the operational concepts that an offensive campaign would require. This was preparation without execution, the loading of a weapon that the political leadership was not yet willing to fire.

Reform in this period extended well beyond the dramatic question of offensive capability. India used the years after 2008 to address a structural weakness that the attacks had exposed with brutal clarity, the failure to fuse information held by different agencies into a single coherent picture. Warnings had existed before the Mumbai assault. They had sat in separate files, held by separate organizations, never assembled by anyone with both the authority and the mandate to act. The reforms that followed created new bodies and new procedures intended to force that fusion, to ensure that a fragment of coastal intelligence and a fragment of communications intelligence and a fragment of human reporting would in future be read together rather than apart. Whether those reforms fully succeeded is debatable, and later failures suggest the problem was never entirely solved. But the effort matters for this history because a service that learns to fuse defensive intelligence more effectively also, as a byproduct, builds a more complete and more actionable picture of its adversary. Machinery built to prevent the next attack was also machinery that sharpened the targeting picture, and the dual use was, once again, unintended. The pattern noted in the third era repeated itself in the fourth. Capability accumulated for defense became, without any single decision to make it so, capability available for offense.

Why did the authorization not come immediately, in the white heat of anger after Mumbai? The answer reveals something important about the Indian system and its brake. The government in office in 2008 considered, in the days after the attacks, a range of retaliatory options, including strikes against militant infrastructure inside Pakistan. It chose restraint. The reasoning combined several factors. The danger of escalation between two nuclear-armed states weighed heavily. The global financial crisis was unfolding, and the international environment counseled caution. India also calculated that restraint itself carried a diplomatic dividend, that a India which absorbed a blow without lashing out would accumulate international sympathy and pressure on Pakistan. That calculation was not foolish, and it was consistent with the country’s long self-image. But it left a residue of frustration in the security establishment and in a section of the public, a sense that restraint had become indistinguishable from impotence. That frustration was the political fuel that the next government would burn.

A political shift came mid-decade. A new government, elected in 2014, arrived with a declared willingness to abandon the restraint that had governed Indian policy since Mumbai. The shift was not instantaneous, but it was real, and it announced itself through conventional military signaling before it manifested in covert action. After a militant attack on an Indian army base in 2016, India publicly conducted what it called surgical strikes across the line in Kashmir, a calibrated raid intended above all as a statement that the old restraint was over. After the suicide bombing of a security convoy in Pulwama in 2019, India sent aircraft across the international border to strike a target inside Pakistan proper, the first such crossing since the 1971 war. These were conventional military operations, not covert intelligence operations, and they belong to a different history. But they matter here because they signaled, unmistakably, that the brake had been released. A government willing to send fighter jets across the border was a government willing, at last, to authorize the quieter and more deniable work that the intelligence service had spent a decade preparing to perform.

The fourth era, then, is the era of the loaded weapon. It produced no confirmed lethal operation inside Pakistan, but it produced the policy consensus, the institutional capability, and the political authorization that the fifth era would require. By the end of the decade, every precondition for an offensive campaign was in place. The networks existed, built across the third era for defensive purposes. The targeting knowledge existed, the files on the militant leadership compiled over many years. The policy will existed, forged in the fury after Mumbai. And the political authorization existed, signaled by a government that had already crossed the border with its air force. What happened next is the campaign that this entire body of work exists to examine.

The Shadow War Era

Around 2021 the fifth and current era begins, defined by a pattern of killings inside Pakistan that, taken together, constitute the most lethal phase of Indian intelligence activity against its neighbor in the agency’s history. This is the shadow war, and the central argument of this article is that it is not a break from everything described above but the culmination of it.

The pattern is by now familiar to anyone who has followed the reporting. Across Pakistani cities, in Lahore, in Karachi, in Rawalpindi, in towns of Punjab and in Pakistan-administered Kashmir, men wanted by India for their role in terrorism have been shot dead by unidentified assailants. The assailants typically arrive on motorcycles, fire at close range, and disappear. Some of the dead were prominent. A man involved in the 1999 hijacking of an Indian airliner was killed in Karachi. A Jaish-e-Mohammed figure long sought by India for the 2016 Pathankot attack was shot in Sialkot. A Hizbul Mujahideen commander was killed in Rawalpindi. A senior Khalistani separatist leader was gunned down in Lahore. The roster of the dead reads like a list drawn directly from India’s most-wanted register, and the geographic spread reads like a map of Pakistan’s terror safe haven network.

What is known about the mechanics of the campaign comes largely from Pakistani investigators and from foreign reporting, and it must be handled with the caution that all such material deserves. Pakistani sources have described, and a major British newspaper investigation in 2024 reported, an operational model that relies on layers of distance. According to these accounts, the actual shooters are frequently local criminals or impoverished men recruited for money rather than ideology, hired through intermediaries who never reveal the ultimate sponsor. Payment, the accounts suggest, often moves through the Gulf, with the United Arab Emirates functioning as a financial and logistical waypoint. Weapons are sourced locally or from the regional black market rather than carried across a border. The design of this model, if the accounts are accurate, is the design of deniability. No Indian officer is anywhere near the killing. The shooter does not know who paid him. The money has crossed multiple jurisdictions. The result is a campaign that produces a body and a pattern but no evidence that survives contact with a courtroom.

A subsequent American newspaper investigation described a comparable structure, reporting that the campaign relied on hired assassins and weapons sourced through Afghanistan, and Pakistani investigators have tied specific 2023 and 2024 killings, including a death in Pakistan-administered Kashmir and a shooting in Sialkot, to the same alleged network. The number of killings attributed to the campaign varies by source and by the period counted, with figures ranging from a handful in the early years to a larger total once the tempo increased. Precision on the exact count is impossible, both because the campaign is unacknowledged and because not every killing of a militant in Pakistan is necessarily the work of a foreign hand. Pakistan is a violent country with its own sectarian feuds, criminal score-settling, and internal militant rivalries. Some deaths attributed to India may belong to those other categories. The honest analyst counts the pattern, not the individual case, because the pattern is what carries the evidentiary weight.

A word is owed here about the named cases, because responsible analysis requires both acknowledging them and refusing to treat any single one as proven. Pakistani investigators have publicly attributed specific killings to the alleged network, and foreign reporting has repeated and in places corroborated those attributions. A figure connected to a notorious aircraft hijacking died in Karachi. A militant leader long sought for an attack on an Indian air base died in a town near the eastern frontier. A separatist organizer died in a major Punjabi city. Each of these deaths fits the pattern, and the accumulation of cases that fit the pattern is what makes the pattern credible. But each case, taken alone, also admits other explanations, because the men in question had enemies far beyond any foreign service, and a country with as much internal violence as Pakistan produces a steady supply of deaths that are genuinely difficult to attribute with confidence. The discipline this article asks of its reader is to hold the cases lightly as individual facts and firmly as a collective pattern. The pattern is the evidence. The single case is only ever a data point within it. This is an uncomfortable epistemological position, because it means the analysis can be confident about the whole while remaining uncertain about every part, but discomfort is not the same as incoherence, and for covert subjects it is frequently the only honest place to stand.

Necessary to state clearly is what India says about all of this, because India’s position is part of the factual record. New Delhi denies everything. The Ministry of External Affairs has described the foreign reporting as false and malicious anti-India propaganda. Senior officials have stated that extraterritorial killing is not the policy of the Indian government. No Indian official has ever confirmed a single operation, claimed a single death, or acknowledged the campaign in any form. This absolute denial is itself a piece of evidence, though not in the direction a casual reader might assume. A government that wished to deter terrorism through fear might be expected to claim its strikes, to let the militant leadership know that the hand reaching for them was Indian. The complete refusal to claim anything suggests either that the campaign is not what the allegations describe, or that deniability is valued so highly that the deterrent benefit of acknowledgment is deliberately forgone. The analytical section below returns to this puzzle in detail.

Set the denial aside for a moment and consider the pattern on its own terms, because the pattern is what connects this era to the four that preceded it. Every element of the shadow war, as it is described, draws on an inheritance. The targeting depends on knowing exactly who the militant leaders are, where they live, what their routines are, which mosque they attend and at what hour. That knowledge is the residue of the third era, the long defensive penetration of the militant ecosystem through the 1990s and 2000s. The willingness to act offensively rather than to file another report is the policy consensus of the fourth era, forged after Mumbai. The use of local proxies kept at arm’s length, criminals and intermediaries rather than Indian officers, is a refinement of the indirect model that goes all the way back to the Mukti Bahini, the recognition learned in 1971 that the most effective and most deniable operations are conducted through local hands. Even the geographic knowledge, the familiarity with Pakistani cities and provinces, has roots in the contacts and the mapping of earlier eras. The shadow war is not an invention. It is an inheritance, assembled.

There is a refinement of the inheritance worth naming precisely. The 1971 model used a proxy that shared India’s strategic goal. The Mukti Bahini wanted what India wanted, the independence of Bangladesh, and so the proxy could be trusted with strategy and direction. The current model, as described, uses proxies who share nothing with India at all. A criminal hired through an intermediary in a Pakistani city has no ideological investment, no knowledge of the larger purpose, and no loyalty beyond the payment. This is not the 1971 model repeated. It is the 1971 model stripped down to its safest essence. The Sri Lankan disaster taught India that an ideologically motivated proxy develops its own agenda and becomes uncontrollable. The shadow war’s answer, if the allegations are accurate, is to use a proxy with no agenda at all, a hired hand who can be discarded after a single use. The evolution from the Mukti Bahini to the anonymous motorcycle shooter is the evolution of an institution learning, across fifty years, how to make the indirect method completely deniable and completely disposable.

One more feature of the current era must be noted, and it concerns the convergence of covert and conventional pressure. In April 2025, a massacre of tourists at Pahalgam in Indian Kashmir killed twenty-six people and triggered the most serious military confrontation between the two states in decades. The Indian conventional response, conducted the following month, involved strikes on targets inside Pakistan and a short, sharp military exchange. The covert campaign and the conventional campaign are distinct, run by different parts of the Indian state, governed by different logics. But they now run in parallel, and their parallel operation marks something genuinely new. For most of the history described in this article, India’s covert hand and India’s military hand were not coordinated instruments of a single strategy. They are becoming one. The current campaign of targeted killings continued, by most accounts, after the conventional confrontation subsided, which suggests a covert track that is now a permanent feature of Indian policy rather than a temporary response to a specific provocation. The overall shape of this campaign, its scope, its tempo, and its strategic logic, is the subject of the shadow war’s foundational analysis.

Key Figures

A history of an institution risks becoming a history of no one, an abstract account of eras and doctrines with no human face. The figures below are not the only people who matter to this story, and in the nature of the trade many of the people who matter most will never be named. But each of these individuals illuminates something essential about how Indian intelligence came to operate inside Pakistan.

Rameshwar Nath Kao

The founding chief of the Research and Analysis Wing is the indispensable figure, because the institution he built carried his imprint for decades. Kao was a careful, cultured, intensely private man, and the agency he created reflected those qualities. He believed in patient construction over flashy action, in the slow building of networks and the careful recruitment of officers, in an institution that would mature into a peer of the great Western and Israeli services rather than one that would seek headlines in its infancy. His defining triumph, the Bangladesh operation of 1971, demonstrated the indirect model at its most effective, the cultivation and direction of a local resistance rather than the deployment of Indian shooters. Kao’s caution is as important as his ambition. He understood that the Bangladesh success owed much to circumstance, and he resisted the temptation to believe the model could be reproduced at will. The institutional personality he established, ambitious in its goals but conservative in its methods and deeply protective of deniability, is visible in the shadow war of the present, which is conducted in exactly the patient, indirect, and unacknowledged style that Kao would have recognized.

Kulbhushan Jadhav

No single individual better illustrates the perils and the murk of Indian intelligence activity inside Pakistan than the man Pakistan arrested in 2016 and identified as a serving Indian intelligence operative. Pakistan claimed he had been captured inside Balochistan and presented him as proof of a vast Indian covert campaign to destabilize that province. India responded that he was a former naval officer running a legitimate business in Iran, that he had been abducted, and that the Pakistani account was a fabrication. The case became a long international legal dispute, with the matter taken to the International Court of Justice, which ruled on questions of consular access without resolving the underlying factual dispute about what the man had actually been doing. The case matters to this history for two reasons. First, it is the closest thing to a public, named test of the Balochistan allegations, and the fact that it remains genuinely contested, that a reasonable observer cannot be certain of the truth, captures the epistemological problem at the center of this entire subject. Second, the case demonstrates the cost of exposure. Whatever the truth, an alleged Indian operative spent years in Pakistani custody, a reminder that the trade described in this article is conducted by individuals who pay, when operations fail, with their freedom or their lives.

Ajit Doval

The figure most associated, in the public imagination, with India’s offensive turn is the former intelligence officer who became the country’s national security adviser in 2014. His own operational career, much of it conducted decades earlier, included extended undercover work, and his reputation as a practitioner of hard intelligence preceded his rise to the senior policy role. As national security adviser, he became the public face of a doctrine that emphasized offense over restraint, the willingness to impose costs rather than absorb them. It would be a mistake to attribute the entire shadow war to one man, and the structural account in this article has deliberately emphasized that the campaign rests on institutional inheritances rather than individual genius. But the elevation of a hard-line intelligence professional to the center of national security policy in 2014 was a signal, and it belongs in any honest account of why the brake was released when it was. The doctrinal framework he came to personify is examined in India’s covert operations doctrine.

The Anonymous Operatives

Final figures in this account have no names, and that namelessness is the point. Every era described above was carried, in its actual execution, by individuals whose identities will never be public. The officers who ran agents into East Pakistan in 1971, the case officers who managed sources inside the militant pipeline through the long defensive decades, the handlers who, if the allegations are true, recruit and direct the intermediaries of the present campaign, all of them work and will continue to work in anonymity. So do the assets, the Pakistani nationals who, for money or conviction or coercion, provide the information and perform the tasks that the campaign requires. The shadow war, as it is described, depends on local hands, on men inside Pakistan recruited to watch a target or to pull a trigger. Those men are not Indian officers. They are the modern equivalent of the indirect instrument that has defined Indian operations since the Mukti Bahini, and the human reality of the campaign is as much their story as it is the story of any chief or adviser. It is also worth remembering that the anonymity protects them only while the operation succeeds. When an operation fails and a shooter is caught, the anonymity collapses, the intermediary chain becomes a chain of evidence, and the disposable proxy discovers exactly how disposable he was.

Consequences and Impact

Consequences of five decades of Indian intelligence activity inside Pakistan are not easily summarized, because they operate at several levels and because the deepest of them are still unfolding. But several can be stated with reasonable confidence.

The first consequence is the hardening of the militant safe haven into a contested space. For decades, the cities of Pakistan functioned as a genuine sanctuary for the men India sought. A militant commander could live in Lahore or Rawalpindi with a degree of openness, attend a mosque, receive visitors, move with relative freedom, secure in the knowledge that the Indian state could reach him only through a legal process that Pakistan would never permit to succeed. The shadow war, if the allegations are accurate, has ended that security. The sanctuary still exists, but it is no longer safe. A commander on India’s list must now consider that the motorcycle pulling up beside him may carry his death. This psychological shift, the conversion of sanctuary into a place of fear, is arguably the campaign’s most significant effect, more significant than any individual death, because it imposes a permanent operational cost on the entire militant ecosystem. A leader who must constantly vary his route, screen his visitors, and limit his public exposure is a leader who has less time and less freedom to plan operations against India. The deterrent effect, if the campaign has one, runs through this drain on the adversary’s operational tempo.

That deterrent claim, however, deserves a skeptical footnote, because the relationship between fear and behavior is rarely as simple as a planner would wish. A militant leadership that feels hunted may indeed plan fewer operations, distracted by the daily work of survival, forced to spend on its own security the energy it would otherwise spend on attacks. But it may equally harden, conclude that it has little left to lose, and accelerate. Fear can deter, and fear can also enrage, and which effect dominates depends on factors that no foreign service fully controls, including the ideology of the targeted movement and the depth of support it draws from the state that shelters it. The honest position is that the campaign almost certainly imposes an operational cost, that the cost is real and probably significant, and that whether that cost translates into measurably fewer attacks on Indian soil is a separate question the available evidence cannot yet answer. Imposing a cost and achieving deterrence are not the same accomplishment, and conflating them is one of the more common errors in the public discussion of this subject. A campaign can be operationally successful, can eliminate the men it sets out to eliminate, and can still fail at the strategic task of making the sponsoring country safer. Success at the tactical level and success at the strategic level are different tests, and only the first of them can be observed from a body count.

A second consequence is reputational and diplomatic, and here the impact runs in both directions. India has acquired, in the eyes of some observers, a reputation as a state willing to conduct extraterritorial killings, a reputation that sits uneasily with the country’s long self-presentation as a responsible status quo power. Allegations of Indian operations in Canada and the United States, separate from the Pakistan campaign, have compounded this. At the same time, India’s absolute deniability has limited the diplomatic damage. Because nothing is confirmed, foreign governments are not forced to respond to an acknowledged Indian policy, and the cost of the campaign in formal diplomatic terms has been contained. Pakistan, for its part, has gained a grievance it can present to the world, but the grievance is complicated by Pakistan’s own long record as a sponsor of militancy, which makes its appeals for international sympathy less effective than they might otherwise be. A state that sheltered the planners of mass-casualty attacks for decades finds it difficult to command outrage when those planners begin to die.

Convergence, already mentioned, is the third consequence, the merging of India’s covert and conventional instruments into something closer to a single coordinated strategy. For most of the history described here, the agency’s work inside Pakistan was separate from the military’s planning, a distinct track with its own logic. The events of the present decade, the surgical strikes, the cross-border airstrikes, and the parallel covert campaign, suggest a state that is learning to use its full spectrum of capability in a coordinated way. This is a significant strategic development, and its consequences for the stability of the relationship between two nuclear-armed states are not yet clear and not entirely reassuring. A relationship in which both the covert and the conventional thresholds have been lowered is a relationship with fewer firebreaks, and fewer firebreaks mean a higher risk that a future crisis escalates beyond the control of either capital.

A fourth consequence belongs to Pakistan’s internal politics and security. The shadow war has exposed the limits of the Pakistani security state’s ability to protect even its own clients. An establishment that built and sheltered militant organizations as instruments of policy now finds that it cannot keep those instruments alive. This is a humiliation for the Pakistani military and intelligence establishment, and humiliations of this kind have consequences, though what those consequences will be remains uncertain. It may drive a reckoning with the policy of sponsoring militancy, a recognition that the proxies have become a liability rather than an asset. It may, alternatively, drive a hardening, a determination to answer the Indian campaign in kind. The history described in this article offers no confident prediction, only the observation that the contest has entered a more dangerous phase, and that a security establishment which feels exposed is a security establishment under pressure to demonstrate that it can still strike back.

A fifth consequence is less tangible but no less real. The shadow war has changed how the entire region thinks about sanctuary and reach. For seventy years, the implicit rule of the India-Pakistan contest was that the international border, and certainly the depth of Pakistani territory, marked a limit. Militancy crossed the border, but the men who directed it sat safely behind it. The campaign of the present decade has erased that limit, or appears to have erased it, and the erasure is itself a strategic fact independent of any individual killing. Other states watch. Other intelligence services study the model. A world in which a determined state can reach into the cities of its rival and eliminate individuals at will, while denying everything, is a different and more dangerous world than the one that existed before, and the shadow war has contributed to bringing that world into being.

Analytical Debate

The central analytical question raised by this history is the one stated at the outset. How extensive, in truth, are Indian intelligence operations inside Pakistan? On this question, two narratives compete, and neither is honest.

Islamabad’s narrative describes a vast and pervasive Indian penetration. In this account, the external intelligence service has agents in every Pakistani city, sleeper cells in every province, a hidden hand behind not only the targeted killings but behind separatist violence in Balochistan, behind unrest in the tribal areas, behind a wide range of the country’s misfortunes. This narrative is politically useful inside Pakistan, because it converts genuine domestic failures, the poverty and marginalization that fuel the Baloch insurgency, the sectarian violence rooted in Pakistan’s own social fractures, into the product of a foreign conspiracy. It absolves the state. It also serves a diplomatic purpose, allowing Pakistan to present itself as a victim of Indian aggression rather than as the long-time sponsor of militancy that the historical record shows it to be. Because the narrative is so useful, it is systematically exaggerated, and a careful observer should discount it heavily.

The Indian narrative is the mirror image, and it is equally dishonest in its own way. New Delhi describes nothing at all. There are no operations, no campaign, no extraterritorial killings, only false and malicious propaganda. This narrative is also politically useful. It allows India to deny a campaign that, conducted in the open, would invite international censure and complicate the country’s relationships with Western partners. It preserves the self-image of the responsible democracy. The trouble with the Indian narrative is that the pattern of dead men is real. The bodies are real. The geographic and methodological consistency of the killings is real, documented by Pakistani investigators and by serious foreign journalism. A blanket denial cannot account for a pattern this consistent, and so the Indian narrative, like the Pakistani one, fails the test of honesty, though it fails in the opposite direction.

Truth, as this article has argued throughout, lies in the contested middle. The most plausible assessment runs as follows. India almost certainly does conduct intelligence operations inside Pakistan, and almost certainly is responsible for at least a significant portion of the targeted killings attributed to it, because the pattern is too consistent and the targeting too precise to be explained otherwise. But the scale is almost certainly far smaller than the Pakistani narrative claims. There is no vast army of Indian agents. There is, more likely, a focused campaign, run through layers of intermediaries, directed at a specific and limited list of wanted men, designed above all for deniability. The campaign is real, but it is surgical, not pervasive. It is a scalpel, and Pakistan describes a flood.

How does one adjudicate between these positions without access to the secret files of either government? The method is the same method a careful historian uses for any covert subject. One examines the pattern rather than the individual case, because a pattern is harder to fake and harder to explain away than a single death. One weighs the incentives of every source, discounting the Pakistani official who gains from exaggeration and the Indian official who gains from denial. One looks for the convergence of independent reporting, treating a finding more seriously when a British investigation and an American investigation, working separately, describe a similar structure. And one applies the test of coherence, asking whether the alleged campaign fits the institutional history. On that last test, the shadow war fits the history of Indian intelligence almost perfectly. It is patient, indirect, deniable, and surgical, exactly the qualities the institution has prized since Kao founded it. A campaign with those characteristics, attributed to an institution with those characteristics, supported by convergent independent reporting and a consistent pattern of evidence, is a campaign a careful analyst can attribute with reasonable confidence, even without a confession.

It is useful, as well, to set the Indian case beside the broader record of states that conduct deniable killings abroad, because the comparison disciplines the analysis. Several governments, across several decades, have run programs of extraterritorial lethal action against people they regarded as terrorists or traitors. Those programs share a recognizable set of features. They rely on layers of intermediary distance. They are officially denied for as long as denial remains sustainable. They are justified, internally, by the claim that no other instrument could reach the target. And they tend, over time, to expand their working definition of a legitimate target, because the machinery, once built and once proven, seeks further work. The alleged Indian campaign fits this comparative template closely, which is a further reason, beyond the pattern and the institutional coherence, to treat the attribution as credible. The comparison also carries a warning that the earlier sections of this article have approached from other directions. The histories of such programs elsewhere include not only operational successes but also the slow erosion of the original limits, the quiet drift from a narrow list of genuine threats toward a broader and more political set of targets. Whether the Indian effort will resist that drift is unknowable from the outside, but the comparative record suggests the pressure toward it is real, and that pressure is one of the things a thoughtful observer should watch for in the years ahead.

This assessment leaves the puzzle of the absolute denial, and the puzzle is worth confronting directly. Why, if India is conducting the campaign, does it not claim the deterrent benefit of acknowledgment? Three explanations are plausible, and they are not mutually exclusive. The first is that deniability is genuinely valued above deterrence, that the Indian state has calculated that the diplomatic protection afforded by silence is worth more than the additional fear that a claimed campaign would generate. The second is that the targeting itself, the pattern of who dies and where, communicates the message to the intended audience without any need for a formal claim. The militant leadership of Pakistan does not need a press release to understand who is killing its commanders. The third explanation is the most sobering, that denial preserves the option of escalation control, that an unacknowledged campaign can be quietly expanded or quietly wound down without the political rigidity that a public commitment would impose. Whatever the correct explanation, the denial should not be read, as a careless observer might read it, as evidence that the campaign does not exist. It is better understood as a feature of the campaign’s design.

A final element of the debate concerns the question of sustainability, and here this history connects to a larger argument. A campaign of targeted killing with no defined endpoint is a campaign that risks becoming permanent, and permanent campaigns eventually produce failures, exposures, and unintended consequences. The shadow war has been, by the available evidence, operationally successful so far. But the history recounted in this article includes the cautionary tale of an earlier era, the proxy operations that produced leverage and then morally fraught entanglement, and the even harder lesson, drawn from a different theater, of a proxy relationship that ended in disaster. The agency’s institutional memory contains these warnings. Whether the current campaign has a theory of its own conclusion, a clear sense of what condition would allow it to stop, or whether it will simply continue until something forces it to stop, is the open question that this history cannot answer. The danger of an open-ended campaign is not that it fails on any given day. The danger is that, lacking an endpoint, it accumulates risk across years until a single exposure, a single captured shooter who can be traced, a single operation that kills the wrong person, converts a quiet success into a public crisis.

Why It Still Matters

The history of Indian intelligence operations inside Pakistan is not a closed file. It is the living foundation of a campaign that continues as these words are written, and understanding the history is the only way to understand the campaign correctly.

A single most important reason this history matters is the one this article has pressed from its opening paragraph. The shadow war is routinely misunderstood as something new, a doctrine invented in 2021, a sudden departure from Indian behavior. That misunderstanding leads to bad analysis. It leads observers to expect that the campaign could be ended as suddenly as it apparently began, that it is a policy choice easily reversed, that it lacks deep roots. The five-era account presented here suggests the opposite. The campaign rests on capabilities, networks, knowledge, and doctrinal lessons accumulated across half a century. The targeting files were built in the defensive decades. The willingness to act offensively was forged after Mumbai. The indirect method, the use of local hands, descends in a direct line from the Mukti Bahini. A campaign with roots this deep is not a passing phase. It is a mature instrument of Indian statecraft, and it should be analyzed as one.

This history matters, second, because it clarifies what the campaign can and cannot achieve. India’s intelligence service has, over fifty years, repeatedly demonstrated that it can reach inside Pakistan and produce an effect. It dismembered the rival state in 1971. It applied leverage in the 1980s. It penetrated the militant ecosystem defensively for two decades. It now, if the allegations hold, eliminates wanted men at will. But the same history shows the limits. The 1971 success depended on circumstances India did not create. The proxy operations of the 1980s produced leverage, not victory. The defensive decades were marked by real failures alongside the successes. Intelligence operations, however skillful, have never resolved the underlying conflict between the two states, and there is no reason to believe the shadow war will resolve it either. The campaign imposes costs. It does not deliver peace, and a policy that mistakes the imposition of costs for a path to resolution is a policy that will be disappointed.

There is a further reason this distinction between cost and resolution matters, and it concerns the politics inside India itself. A campaign that produces visible results, dead enemies, a tangible sense of reach and capability, is politically rewarding in a way that the patient and invisible work of diplomacy can never be. A government can point to the weakening of a hostile network and invite its public to feel that something decisive has been done. It cannot easily point to a crisis that never happened because a negotiation quietly succeeded. This asymmetry creates a subtle incentive to prefer the instrument that performs well in public over the instrument that might actually reduce the underlying danger. The risk is not that the covert campaign is useless, because it plainly is not. The risk is that its very effectiveness as a source of national satisfaction crowds out the harder, slower, and far less dramatic work that a genuine settlement of the contest would require. A tool that works can become a substitute for a strategy, and the history traced in this article offers more than one example of exactly that substitution, the proxy operations that produced leverage and then stood in for a policy, the defensive decades that managed a threat without ever resolving it. The shadow war risks becoming the latest entry in that pattern, an instrument so satisfying to wield that the question of what it is ultimately for goes quietly unasked.

A third reason this history matters is that the relationship between the two states has entered a phase more dangerous than any since the conflicts of the last century. The convergence of covert and conventional instruments, the demonstrated willingness on both sides to cross thresholds that were once respected, the events of 2025, all of this points to a contest that is escalating rather than stabilizing. Two nuclear-armed states are testing how far the shadow war and the open confrontation can be pushed, and they are doing so without a clear understanding, on either side, of where the limits lie. A history that explains how the two countries arrived at this point is not an academic exercise. It is the necessary background for anyone trying to assess where the contest goes next, and the assessment is not encouraging.

This history matters, finally, because it offers a corrective to the way intelligence is usually narrated. The popular account of espionage is a story of brilliant operations and decisive moments. The real story, as the five eras above demonstrate, is a story of slow accumulation, of capability built quietly across decades for one purpose and then turned, when the politics permitted, to another. The shadow war did not spring from a single decision. It was assembled, year by year, from a war, a set of contested proxy operations, two decades of patient defensive work, and a policy revolution born of catastrophe. To understand any covert campaign, the analyst must resist the drama and trace the accumulation. That is the discipline this article has tried to model.

For the reader who has followed this account from the war that created Bangladesh to the killings of the present decade, a final observation is this. The motorcycle that pulls up beside a wanted man in a Pakistani city, the two shots, the disappearance into traffic, all of it can look like a thing without history, a clean and modern operation severed from any past. It is the opposite. Behind that motorcycle stands fifty years of institution-building, of networks cultivated and lost and rebuilt, of lessons learned in triumph and in disaster, of a brake applied and released and applied again. The shadow war is the inheritance of everything described here. To see it clearly is to see the long history behind it, and the long history is the point. A reader who carries away only one idea from this account should carry away that one, because it is the idea that corrects the most common error and unlocks the most accurate analysis. Nothing in the trade of intelligence arrives without a past. The operation that looks newest is usually the oldest, an accumulation of decades wearing the costume of a fresh decision, and the analyst who learns to look past the costume will understand the contest between these two states far better than the analyst who is content to be surprised.

Frequently Asked Questions

What is the history of RAW operations inside Pakistan?

Indian intelligence activity inside Pakistan stretches back to the founding of the Research and Analysis Wing in 1968, but it is best understood as five distinct eras. The first was the 1971 war, when the agency organized and directed the Bengali resistance that produced Bangladesh. The second, in the 1980s, involved alleged covert support for separatist movements in Sindh and Balochistan as a response to Pakistan’s sponsorship of insurgency in Indian Punjab. The third, running through the 1990s and 2000s, was a defensive era of collection and counter-intelligence focused on the Kashmir insurgency. The fourth, after the 2008 Mumbai attacks, saw a policy turn toward offensive thinking and the building of offensive capability. The fifth and current era is the shadow war, the campaign of targeted killings that began around 2021.

Has RAW operated in Balochistan?

This is among the most contested questions in the entire subject. Pakistan alleges extensive Indian involvement in Balochistan, presenting it as the central evidence of a destabilization campaign. India denies any role and attributes the Baloch insurgency entirely to Pakistan’s own domestic failures. The honest assessment is that the Baloch movement has authentic and longstanding causes rooted in poverty, political marginalization, resource appropriation, and the heavy hand of the security forces, and that attributing it to foreign manipulation is both inaccurate and an insult to the Baloch. At the same time, it is plausible that Indian intelligence established contacts in the province during the 1980s and may retain some presence. The likely truth is a modest Indian footprint dramatically exaggerated by Pakistani officials for political and diplomatic reasons.

How extensive is RAW’s intelligence network in Pakistan?

Less extensive than Pakistan claims and more extensive than India admits. Pakistan describes a pervasive penetration, agents in every city, sleeper cells everywhere, a hidden hand behind every misfortune. India describes nothing at all. The most plausible assessment is a focused and limited network, built over decades, capable of supporting a surgical campaign against a specific list of targets, but nothing resembling the vast army of agents that Pakistani officials describe. The exaggeration serves Pakistan politically, because it converts domestic problems into a foreign conspiracy and shifts responsibility away from the state.

Does RAW have agents in Pakistani cities?

The pattern of targeted killings strongly suggests that the campaign depends on local presence, on people inside Pakistani cities who can identify a target, track his routine, and either perform or facilitate the operation. According to Pakistani investigators and foreign reporting, those local hands are frequently criminals or impoverished men recruited for money through intermediaries, rather than Indian officers. So the answer is that the campaign relies on assets inside Pakistani cities, but those assets are largely local recruits kept at a distance from any Indian official, which is the design that makes the operations deniable.

What operations did RAW conduct in Pakistan before the shadow war?

The most significant was the 1971 operation that produced Bangladesh, when the agency organized, trained, armed, and directed the Mukti Bahini and provided intelligence preparation for the conventional military campaign. The 1980s saw the alleged Sindh and Balochistan operations conducted as leverage against Pakistan’s Punjab proxy war. The 1990s and 2000s were dominated by defensive intelligence work, the penetration of militant organizations and the interception of plots, rather than dramatic offensive operations. No confirmed lethal operation inside Pakistan is documented before the current era.

How does Pakistan describe RAW’s presence?

Pakistan describes the Indian intelligence presence in the most expansive possible terms, as a pervasive and malign penetration responsible for the targeted killings, for separatist violence, for sectarian unrest, and for a broad range of the country’s troubles. This description is politically useful inside Pakistan because it shifts blame for domestic failures onto a foreign enemy, and it is diplomatically useful because it allows Pakistan to present itself as a victim. The description should be read with heavy skepticism, because the incentive to exaggerate is overwhelming.

Is the shadow war RAW’s first lethal operation in Pakistan?

In the sense of a sustained campaign of targeted killings, yes, the shadow war appears to be without precedent in the agency’s history. Earlier eras involved war, proxy support, leverage, and defensive penetration, but not a documented pattern of assassinations. This is what makes the current era genuinely distinct. It is also what this article cautions against misreading, because while the lethal campaign is new, the capabilities it depends on are not.

What infrastructure did previous operations build for the shadow war?

Three inheritances are essential. The defensive decades of the 1990s and 2000s built an intimate knowledge of the militant ecosystem, the leaders, the safe houses, the couriers, the routines, knowledge acquired to stop attacks on Indian soil but later usable for targeting. The era after the 2008 Mumbai attacks built the policy consensus that offensive action was necessary and the institutional capability to conduct it. And the indirect method, the reliance on local hands rather than Indian officers, descends from the 1971 model of cultivating a proxy. The shadow war assembled these inheritances rather than inventing anything.

Did India support separatists inside Pakistan in the 1980s?

The evidence suggests that India probably did establish contacts with separatist movements in Sindh and Balochistan during the 1980s and probably did provide some level of support, conceived as a mirror response to Pakistan’s sponsorship of the Khalistan insurgency in Indian Punjab. But the evidence is thin and dependent on interested sources, the support was likely modest rather than decisive, and it functioned more as leverage and signaling than as a serious attempt to break up the Pakistani state. The two countries appear to have reached a partial understanding by the early 1990s under which India wound the operation down.

Who is Kulbhushan Jadhav and why does his case matter?

He is a former Indian naval officer whom Pakistan arrested in 2016 and identified as a serving Indian intelligence operative captured inside Balochistan. India said he was running a legitimate business in Iran and had been abducted. The case became a long dispute that reached the International Court of Justice, which ruled on consular access without settling the factual question of what he had actually been doing. His case matters because it is the closest thing to a public, named test of the Balochistan allegations, and the fact that it remains genuinely unresolved captures the central difficulty of the whole subject, the near-impossibility of certainty.

Why does India deny the shadow war so completely?

Several explanations are plausible and they are not mutually exclusive. Deniability may simply be valued above the deterrent benefit of acknowledgment, because silence protects India diplomatically. The targeting itself may communicate the message to the militant leadership without any need for a formal claim. And denial preserves flexibility, allowing the campaign to be quietly expanded or wound down without the rigidity that a public commitment would impose. The absolute denial should not be read as evidence that the campaign does not exist. It is better understood as part of the campaign’s design.

How does RAW keep its operations deniable?

According to Pakistani investigators and foreign reporting, the campaign is built around layers of distance. The actual shooters are often local criminals recruited for money rather than ideology, hired through intermediaries who never reveal the ultimate sponsor. Payment frequently moves through the Gulf, with the United Arab Emirates serving as a financial waypoint, and weapons are sourced locally or regionally rather than carried across a border. The design ensures that no Indian officer is near the killing and that no evidentiary chain survives, which produces a body and a pattern but nothing that would hold up in a court.

Did the 1971 war count as an operation inside Pakistan?

Yes, and it was the most consequential one in the agency’s history. East Pakistan was, at the time, part of the Pakistani state. The agency’s organization and direction of the Bengali resistance, its training camps, its supply of weapons, and its intelligence preparation for the Indian military campaign together constituted a covert operation of extraordinary scale, one that ended with the rival state losing more than half its population. The operation established the agency’s appetite for working inside Pakistani territory and proved that such work could redraw a map.

How did the 2008 Mumbai attacks change RAW’s posture toward Pakistan?

The Mumbai attacks did not change capability overnight, but they changed the conversation about what capability should be used for. The attacks exposed the failure of defensive intelligence and, in their aftermath, the failure of legal and diplomatic pressure to deliver any consequence to the planners. A section of the Indian strategic community concluded that if the men who planned mass-casualty attacks were to face consequences, India would have to deliver them directly. This was the intellectual origin of the offensive turn, a turn in policy thinking that preceded by several years the offensive operations themselves.

Is RAW responsible for the targeted killings in Pakistan?

India denies all responsibility. But the pattern of dead men is real, the targeting is precise, and the geographic and methodological consistency is documented by Pakistani investigators and serious foreign journalism. The most plausible assessment is that Indian intelligence is responsible for at least a significant portion of the killings, because no other explanation accounts for a pattern this consistent. The campaign is real, though it is surgical and limited rather than the pervasive penetration Pakistan describes. Some individual deaths attributed to India may belong to other categories, such as internal militant rivalry or criminal score-settling, which is why the honest analyst weighs the pattern rather than the single case.

How does RAW’s Pakistan record compare to its other operations?

The agency’s signature successes elsewhere include the Bangladesh operation of 1971 and the engineering of Sikkim’s accession in 1975, a regime-change operation accomplished without a casualty. Its most painful failure unfolded in Sri Lanka, where support for a Tamil militant group ended in disaster and in the deaths of Indian soldiers. The Pakistan record is distinct because it is the longest-running of all, spanning the agency’s entire existence, and because it has moved through more phases, from war to proxy support to defensive penetration to the current lethal campaign, than the agency’s involvement in any other country.

Has RAW’s approach to Pakistan changed under the current Indian government?

The evidence suggests it has. The government elected in 2014 arrived with a declared willingness to abandon the restraint that had governed Indian policy since 2008, and that willingness manifested first in conventional signaling, the surgical strikes and the cross-border airstrikes, and then, by most accounts, in the covert campaign that began around 2021. The capability had been built earlier, across the defensive decades and the years after Mumbai. What changed under the current government was the authorization to use it.

Will the shadow war inside Pakistan continue?

The available evidence suggests the campaign is now a permanent feature of Indian policy rather than a temporary response to a specific provocation, because the killings reportedly continued after the conventional military confrontation of 2025 subsided. The deeper question is whether the campaign has a theory of its own conclusion or whether it will simply continue until something forces it to stop. The history of Indian intelligence operations inside Pakistan contains cautionary lessons about open-ended commitments, but whether those lessons will shape the current campaign’s future is unknown.

What is the biggest risk of the current campaign for India?

The biggest risk is exposure. A campaign built entirely on deniability collapses the moment the deniability fails, and the more operations a campaign runs, the higher the cumulative chance that one of them goes wrong. A captured shooter who can be traced back through the intermediary chain, an operation that kills the wrong person, a financial trail that investigators manage to follow, any of these could convert a quiet and successful campaign into a public diplomatic crisis. The history of covert operations is full of programs that succeeded for years and then failed catastrophically in a single exposed case, and an open-ended campaign with no defined endpoint accumulates exactly this kind of risk over time.