concurring in part, dissenting in part.
Considering the evidence on record, and drawing all reasonable inferences in favor *24of the non-moving plaintiffs, I believe the majority judges are incorrect in affirming the grant of summary judgment as to all supervisory defendants. Though a close call, I find there are questions of material fact regarding the supervisory liability of Cruz-Sánchez and Colón-Báez that have improperly been kept from a jury.13
Specifically, I believe there are questions of fact on whether officers Cruz-Sánchez and Colón-Báez were on notice of Pagán’s seriously violent tendencies. I also believe it should be up to a jury to determine whether the failure of these supervisory officers to take any measures to prevent or at least mitigate the grave risk that Pagán posed to the constitutional rights of others is causally related to the shooting death of Cáceres. Accordingly, I respectfully dissent.
I. Background
Review of a district court’s grant of summary judgment is de novo. Euromodas, Inc. v. Zanella, Ltd., 368 F.3d 11, 16 (1st Cir.2004). In conducting this review, we draw all reasonable inferences in favor of the non-moving party. Collazo v. Nicholson, 535 F.3d 41, 44 (1st Cir.2008). “Summary judgment is appropriate where there exists no genuine issue of material fact and the moving party is entitled to judgment as a matter of law.” Id.
A. Pagan’s record and appointment to the TOD and the IU
Prior to fatally shooting Cáceres, Pagán had seven complaints on his PRPD disciplinary record. These included a complaint for insubordination, one for domestic violence, and one for assault on a motorcyclist. The latter was, however, found to be unsubstantiated by the PRPD. The plaintiffs have provided some form of proof for each of the seven complaints.
The plaintiffs’ argument rests in large part on the domestic violence complaint, filed by Pagán’s then girlfriend, which itself encompassed three acts of serious violence and police impropriety by Pagán in 1998. In the first incident, Pagán struck his girlfriend and, while brandishing his firearm, threatened her with death. In the second incident, Pagán visited his girlfriend’s home carrying a gun he had taken from an arrestee. He brandished the gun and swore he would kill the arrestee whom the weapon belonged to. He then stored the weapon at his girlfriend’s apartment for a few days. Finally, after their relationship had ended, Pagán stormed into his ex-girlfriend’s apartment and once again assaulted and threatened her.
The PRPD’s investigation of the domestic violence complaint filed by Pagán’s girlfriend began in 1998. Pagán did not receive a sanction, however, until August 2004, when then Superintendent Agustín Cartagena ordered his expulsion from the force. While the order of expulsion against Pagán was pending, Cruz-Sánchez promoted Pagán to the Humacao Tactical Operations Division (“TOD”). Cruz-Sán-chez was Director of the TOD at the time. The TOD is a specialized team of “elite” officers within the PRPD who receive additional training for particularly sensitive situations. Cruz-Sánchez did not assess Pagán’s disciplinary history before promoting Pagán to the TOD.
In early 2005, defendant Toledo-Dávila took on the job of Superintendent, and in December 2005, he reissued Pagán’s order of expulsion. After Pagán sought internal administrative review of this sanction, Toledo-Dávila reduced his punishment to a *25sixty-day suspension without pay. Colón-Báez served Pagán the suspension papers.
Pagán served his suspension from August 23 to October 22, 2006. At that time, Cruz-Sánchez was still Director of the TOD, with Colón-Báez as Assistant Director. Upon completion of his suspension, Pagán immediately rejoined the TOD. Though PRPD regulations require that officers coming back from suspension be sent initially to Replacement Centers, none of the supervising officials took any action to transition Pagán back into service. In January 2007, shortly after Pa-gán served his suspension, Cruz-Sánchez evaluated Pagán and gave him stellar reviews. He admits he did so without reviewing Pagán’s disciplinary file, or otherwise investigating his disciplinary history. This evaluation was seven months before the shooting of Cáeeres.
A few months later in the summer of 2007, defendant Rivera-Merced, a high ranking PRPD official, sought to create a specialized Impact Unit (“IU”) within the TOD, for intervention in high crime areas. Cruz-Sánchez developed the operational plan for the IU, and Colón-Báez handpicked Pagán for the unit; Pagán was again promoted. Neither Cruz-Sánchez nor Colón-Báez reviewed Pagán’s disciplinary file or otherwise probed his disciplinary background before accepting Pa-gán into the IU.
B. The murder of Cáeeres and the immediate aftermath
I will spare the details of Cáeeres’s execution. Suffice it to say that the shooting was caught on video and the circumstances surrounding it are not in dispute.
Immediately after the murder of Cá-eeres, Pagán and the two other defendant line officers who were present at the shooting left for Ryder Hospital nearby in Humacao. The first to arrive on the scene of Cáceres’s murder was Detective Rodriguez, who is not a party to this suit. After speaking to a number of witnesses, Detective Rodriguez described the incident to Cruz-Sánchez, apparently over the phone or radio. Cruz-Sánchez was the highest ranking officer in the Humacao area that evening. Detective Rodriguez told Cruz-Sánchez that he had no doubt the officers had abused their power. Detective Rodriguez prepared a report of the incident, essentially relaying that Pagán shot Cá-eeres four times while the latter lay on the ground, and expressing his view that the use of deadly force was entirely uncalled for.
Meanwhile, as Detective Rodriguez was investigating the scene of the crime, Cruz-Sánchez joined Pagán at Ryder Hospital. Colón-Báez, even though he was not on duty that evening, also joined Cruz-Sán-chez and Pagán at the hospital.
As Cruz-Sánchez spoke to Pagán and the other defendant line officers for their version of events, other eyewitnesses of the shooting began to arrive at the hospital and offered their own account of what had transpired: that Cáeeres had been shot dead while lying defenseless on the floor. By this time, Detective Rodriguez had already apprised Cruz-Sánchez that an act of police brutality had occurred.
Notwithstanding clearly contradictory accounts from other eyewitnesses, the report that resulted from information retold by Cruz-Sánchez and Colón-Báez, adopted the version of Pagán, his companion officers and that of an Héctor Huertas, the only bystander identified by name in the report, and, coincidentally, the only witness who gave a view of the events that was favorable tó Pagán. As to the numerous accounts unfavorable to the officers, the report merely notes that “several persons, were interviewed at the scene” who *26“provided information that was adverse and against the agents.” A few days later, Cruz-Sánchez himself added information to the report related to the identity of witness Héctor Huertas and noted that the latter’s version coincided in part with that of Pagán. Cruz-Sánchez noted that this corroboration compelled him to believe Pa-gán’s version of the incident: that Cáceres had turned violent, attempted to wrestle his gun from him, and “several shots were fired” in the skirmish that ended in Cá-ceres’s death.
The accuracy of Cruz-Sánehez’s report of the events of August 11, 2007 came under serious doubt shortly thereafter when, thanks to a bystander who video-recorded the incident, the execution of Cá-ceres was aired on the evening news a few days later.
C. PRPD General Order 87-11
PRPD General Order 87-14 (“G.O. 87-14”) requires officials in supervisory positions to examine the personnel file of each and every officer under their supervision. According to G.O. 87-14, supervisors must assess whether an officer in their unit is of violent character or holds the potential to commit civil rights violations. A supervisor must make this independent assessment whether or not there are substantiated complaints against the officer. G.O. 87-14 mandates that prior conduct of, and complaints against, the officer must be assessed in light of the underlying facts of the incident, and not on the ultimate result of the complaint.
II. Discussion
A supervisory official may be found liable under 42 U.S.C. § 1983 for actions of his own that result in violations of constitutional rights by a subordinate. Camilo-Robles v. Hoyos, 151 F.3d 1, 6-7 (1st Cir.1998). A supervisor “may be liable for the foreseeable consequences of such conduct if he would have known of it but for his deliberate indifference or willful blindness.” Maldonado-Denis v. Castillo-Rodríguez, 23 F.3d 576, 582 (1st Cir.1994). To prevail on a theory of deliberate indifference, “a plaintiff must show (1) a grave risk of harm, (2) the defendant’s actual or constructive knowledge of that risk, and (3) his failure to take easily available measures to address the risk.” Figueroa-Torres v. Toledo-Dávila, 232 F.3d 270, 279 (1st Cir.2000) (internal quotation marks omitted). Liability does not attach on a showing of deliberate indifference alone, however; there must be an affirmative link between the subordinate’s misconduct and the action, or inaction, of supervisory officials. See Id. This causal connection “need not take the form of knowing sanction, but may include tacit approval of, acquiescence in, or purposeful disregard of, rights-violating conduct.” Hoyos, 151 F.3d at 7 (citing Maldonado-Denis, 23 F.3d at 582 (explaining that the supervisor must have “had the power and authority to alleviate [the violation]”)).
To be sure, both Cruz-Sánchez and Co-lón-Báez dispute some aspects of plaintiffs’ version of the facts. All the more reason to conclude that, on this record, summary judgment in favor of Cruz-Sán-chez or Colón-Báez was inappropriate.
As to their knowledge — deemed or otherwise — of Pagán’s violent tendencies, there is arguably some dispute as to whether G.O. 87-14 had been suspended, and when exactly this suspension might have happened. Cruz-Sánchez testified, however, that as a supervisor he was responsible for reviewing the disciplinary files of officers under his supervision. This is consistent with the testimony of defendant Superintendent Toledo-Dávila, who stated that all supervisors were charged with the responsibility of review*27ing subordinates’ personnel files. Colón-Báez claims that he had no access to Pa-gán’s disciplinary file, though there is also some dispute as to that. These questions turn on an assessment of credibility and, accordingly, are best left for the jury to answer.
Moving on to the pudding, one defendant at a time, Cruz-Sánchez was a lieutenant and former Director of the TOD, and he personally promoted Pagan to the TOD in 2004. At that time, there was a pending expulsion order against Pagán. Though plaintiffs dispute the claim, Cruz-Sánchez claims he was unaware of the pending expulsion order. Cruz-Sánchez did admit that, though he was required to do so, he did not review Pagán’s disciplinary file. Even assuming that he did not know the nature or the extent of the violations that led to Pagán’s pending expulsion order, that G.O. 84-17 required him to review Pagán’s file — a requirement he admittedly knew of — should have compelled him to review the file. It is clearly a question for the jury whether this circumstance is deemed to put Cruz-Sánchez on notice of Pagán’s violent character.
Later, Cruz-Sánchez again had the opportunity to keep Pagán from a position where he posed a danger to civilians, but remained idle. In October of 2006, upon completion of a sixty day suspension, Cruz-Sánchez took Pagán right back into the TOD.14 He again neglected to review Pagán’s disciplinary file, despite the fact that he was aware that Pagán had just served a considerable sanction. Two to three months later in January 2007, again neglecting to review Pagán’s file or otherwise inquire into his disciplinary history, Cruz-Sánchez evaluated Pagán and gave him stellar reviews, with high marks on the category of self-control. Seven months later, Cáceres lay dead on the ground. In fact, Cruz-Sánchez had perhaps one more opportunity to take action. And it may have come in the Summer of 2007 when Cruz-Sánchez designed the plan for the IU, a specialized squad within the TOD, and Colón-Báez chose its members, and Pagán was allowed to join. But again, no effort was made to investigate Pagán’s disciplinary history.
It is up to the jury to consider Pagán’s return to the TOD and shortly thereafter his promotion to the IU, closely following a suspension, all at the behest of Cruz-Sán-chez, who oversaw these decisions and afforded Pagán glowing reviews, without even glancing at his disciplinary record, much less investigating the reasons for his suspension. Sub par, pro forma evaluations in particular concerned us in Gutierrez-Rodríguez v. Cartagena, 882 F.2d 553 (1st Cir.1989), when the supervisory defendant in that case gave favorable reviews to an officer that had ten complaints against him, including some episodes of violence. Id. at 563, 582.
Colón-Báez presents a more egregious case than Cruz Sánchez. Colón-Báez personally selected Pagán for the IU, knowing full well of Pagán’s violent past; he personally served Pagán with his suspension papers and admitted knowing that the *28sanction was the result of domestic violence. Notwithstanding this fact, Colón-Báez chose not to investigate Pagán’s disciplinary history, which would have uncovered several episodes of misconduct, including a few particularly violent ones, and selected Pagán for the IU on a whim.
The majority contends that there is no evidence of Colón-Báez’s knowledge of Pa-gán’s violent past. Though the extent of Colón-Báez’s knowledge is arguably disputed, there is no question that he was aware that Pagán served a considerable suspension for an episode of domestic violence. Furthermore, had he probed the matter, as he was required to do, he would have discovered that Pagán had battered his ex-girlfriend and threatened her with death while brandishing his firearm.
As to Colón-Báez, a rational jury could conclude that he relied on no criteria for selecting prospective members of the IU, and in Pagán’s case in particular, declined to review his personnel file or assess his disciplinary background. I stress that Co-lón-Báez’s inaction is particularly objectionable given that it is undisputed that he was aware that Pagán faced disciplinary action due to violent conduct. A rational jury could easily conclude that awareness of Pagán’s suspension should have, at the very least, put him on notice of Pagán’s violent character.
The majority also takes the position that review of Pagán’s disciplinary file would not have uncovered much of his violent past. This is so, they contend, because at some point during his tenure as Superintendent, Toledo-Dávila limited the information contained in disciplinary orders by removing the factual details underlying complaints, and listing only the sanction. This, however, does not provide shelter to Cruz-Sánchez or Colón-Báez. A change in the format of disciplinary orders did not relieve either supervisor of his duty to ascertain whether or not officers under their supervision presented a danger to the civil rights of citizens. Both G.O. 87-14 and the testimony of Toledo-Dávila regarding the duty of supervisors to assess their subordinates’ character bolster this proposition. Furthermore, as Cruz-Sán-chez himself admitted, both he and Colón-Báez could have easily accessed whatever information was pertinent to Pagán’s suspension, which they were both aware of, merely by requesting it.
I am concerned by the majority’s view that Pagán’s disciplinary history was not enough to put the supervisory officials on notice that he presented a substantial risk of shooting an arrestee or civilian. Underlying this finding is the notion that, in order for liability to attach on a deliberate indifference theory, our case law requires that supervisory officials be on notice, not merely of the potential for violence on the part of the subordinate, but of the potential of a specific act of violence, in this case, shooting a civilian.
To be sure, the Supreme Court has provided guidance to the effect that there must be warning of a specific kind of injury. See Bd. of Cnty. Comm’rs of Bryan Cnty., Okla. v. Brown, 520 U.S. 397, 412, 117 S.Ct. 1382, 137 L.Ed.2d 626 (1997) (“[A] finding of culpability simply cannot depend on the mere probability that any officer inadequately screened will inflict any constitutional injury. Rather it must depend on a finding that this officer was highly likely to inflict the particular injury suffered by the plaintiff.”) (emphasis omitted). However, if a subordinate’s threats of death by gunfire against another person are not enough to put a supervisor on notice that the subordinate is a prime prospect for engaging in such conduct in the future, is it required that his supervisors wait until the subordinate actually commits such a crime before corrective or *29preventive measures are taken? Such a strenuous standard cannot possibly be the law. In the case of Pagán, after one episode of executing a civilian, it seems obvious now that he is an ideal candidate for supervisory action based on his proven record. For Cáceres, it was one shot too many.
Accordingly, out of an abundance of caution, I reject any reading of the case law that approaches affording defendants one free bite at the apple. Though in an Eighth Amendment context, our ease law has actually disavowed the idea — which would essentially amount to requiring clairvoyance — that too much specifics are required. See Ruiz-Rosa v. Rullan, 485 F.3d 150, 157 (1st Cir.2007) (“[P]roof of deliberate indifference by prison officials does not require evidence that the officials were aware of the risk of a specific harm.”). I think we’ve avoided such a rule for good reason.
I concede that whether the causal connection here is sufficient, is a close question, particularly as to Cruz-Sánchez. I understand it may seem a stretch to some, at first glance, that a few violent episodes in 1998 would somehow be linked to another violent episode in 2007. However, it is in part because this is a difficult question that I believe the majority errs in not allowing the jury to fulfill its traditional function. Young v. City of Providence, 404 F.3d 4, 23 (1st Cir.2005) (“[Questions of proximate cause are generally best left to the jury.”). Though Pagán’s most egregious acts of violence happened years before the murder of Cáceres, the disciplinary proceedings related to those acts did not conclude until eight years later, in October 2006, when Pagán served his suspension only months before the execution.15
A reasonable jury could conclude that Cruz-Sánchez and Colón-Báez had several opportunities throughout these years to take action to prevent harm to civilians on behalf of Pagán, and failed to act up until 2007. Not an unlikely possibility, given that, only months later, he would carry out the murder he had threatened against at least two other people in the past, and took Cáceres’s life.16
III. Conclusion
A jury should have the opportunity to determine whether Cruz-Sánchez and Co-lón-Báez were on notice of the risk of harm Pagán posed to civilians. It should *30also have the occasion to determine whether either defendant should have seized any of the opportunities they had to keep Pagan from acting out and repeating his violent tendencies. I believe a reasonable jury could answer both inquiries in the affirmative. A claim of ignorance cannot shield them from liability. In fact, such a claim might be probative of deliberate indifference. Accordingly, for these reasons, I dissent.
. I agree with the majority's opinion as to the remaining defendants on appeal.
. The majority judges contend that it was Rivera-Merced’s decision to accept Pagán back into the TOD immediately following his suspension, and not Cruz-Sánchez's. Rivera-Merced — who also claims ignorance of Pa-gán’s disciplinary history — as Commander of the Humacao area had the authority to assign Pagán to any position within the Humacao area. There is no evidence, however, that Rivera-Merced actually ordered that Pagán be accepted back on the TOD. More importantly, that Rivera-Merced was higher up the hierarchical ladder did not relieve Cruz-Sán-chez of his own duty, as Director of the TOD, of screening recruits and forming his own assessment of their suitability for service in the TOD.
. Respectfully, I find improbable the majority’s belief that the PRPD handled the complaints against Pagán in a serious manner. The mere fact that eight years elapsed before Pagán received a final sanction makes that procedure seem, frankly, laughable. Moreover, in most jurisdictions I am aware of, assault and battery with a deadly weapon carries a sentence of incarceration. Thus, a sixty day suspension seems quite insubstantial as an administrative sanction for essentially the same conduct.
. The majority cites to Barreto-Rivera v. Medina-Vargos, 168 F.3d 42 (1st Cir.1999), in support of its finding that Pagán’s disciplinary file was not substantial enough to alarm his supervisors that he posed a grave risk of harm. In Barreto-Rivera we found sufficient evidence to conclude that a supervisory official was deliberately indifferent to, among other things, a disciplinary history more extensive than Pagán's. Id. at 49. The principal defendant in Barreto-Rivera had also shot and killed a civilian after a personal altercation. The brunt of our analysis in Barreto-Rivera turned on the question of whether the principal defendant was acting under color of state law. Id. at 46-48. We also found a causal nexus to the supervisory official's omissions because there was a "known history of widespread abuse” as evinced by the principal defendant’s extensive disciplinary record. Id. at 48-49. I note however, that Barreto-Rivera did not set a floor or a ceiling for disciplinary records on the question of what amounts to notice of a risk of harm.