dissenting in part:
I fully agree that there was no error in the failure to hold an evidentiary hearing on Carey’s claim that the voluntariness of his plea was compromised by his ingestion of prescription medications prior to the change of plea hearing. I disagree, however, that the district court permissibly rejected Carey’s conflict of interest claim without an evi-dentiary hearing. For this reason, I respectfully dissent from Part B of the majority opinion.
I.
At the outset, I think it important to state some bedrock legal tenets. First, “the right to counsel is the right to effective assistance of counsel.” McMann v. Richardson, 397 U.S. 759, 771 n. 14, 90 S.Ct. 1441, 1449 n. 14, 25 L.Ed.2d 763 (1970). Second, the right to effective assistance of counsel is always denied where an actual conflict of interest negatively affects a lawyer’s performance. See Cuyler v. Sullivan, 446 U.S. 335, 345-50, 100 S.Ct. 1708, 1716-19, 64 L.Ed.2d 333 (1980). In other words, an actual conflict of interest is a special breed of ineffective assistance *1102which is never harmless and is presumptively prejudicial. Id. at 349-50, 100 S.Ct. at 1718-19. And third, the right to effective assistance of counsel applies at the sentencing stage of a felony case. See Mempa v. Rhay, 389 U.S. 128, 134, 88 S.Ct. 254, 256, 19 L.Ed.2d 336 (1967).
I recognize that the standard by which effectiveness ordinarily is judged may be more lax at the sentencing stage of a noncap-ital case than it is at trial. Cf. Strickland v. Washington, 466 U.S. 668, 686, 104 S.Ct. 2052, 2063, 80 L.Ed.2d 674 (1984) (leaving open the question whether Strickland’s effectiveness definition should apply to counsel’s performance at “an ordinary sentencing, which may involve informal proceedings and standardless discretion in the senteneer, and hence may require a different approach to the definition of constitutionally effective assistance”).4 There can be no doubt, however, that a lawyer whose performance at sentencing was compromised by an actual conflict of interest has not rendered the client the effective assistance of counsel mandated by the Constitution. See United States v. Swartz, 975 F.2d 1042, 1048 (4th Cir.1992) (applying Cuyler at the sentencing stage); United States v. Ziegenhagen, 890 F.2d 937, 939-41 (7th Cir.1989) (same); see also United States v. Green, 680 F.2d 183, 191-205 (D.C.Cir.1982) (Bazelon, J., dissenting), cert. denied, 459 U.S. 1210, 103 S.Ct. 1204, 75 L.Ed.2d 445 (1983).
II.
In light of the foregoing authority, it is clear that the appropriate inquiry here simply is whether Carey’s allegation is sufficient to state a claim that his lawyer had an actual conflict of interest. As the majority concedes, under 28 U.S.C. § 2255, a petitioner is entitled to an evidentiary hearing “[u]nless the motion and the files and records of the case conclusively show that the prisoner is entitled to no relief....” Thus, a petition can be dismissed without a hearing only if the petitioner’s allegations, if credited, would not entitle the petitioner to relief, or “if the allegations cannot be accepted as true because they are contradicted by the record, inherently incredible, or conclusions rather than statements of fact.” United States v. Rodriguez Rodriguez, 929 F.2d 747, 751 (1st Cir.1991); see also Dziurgot, v. Luther, 897 F.2d 1222, 1225 (1st Cir.1990); Mack v. United States, 635 F.2d 20, 26-27 (1st Cir.1980).
Here, Carey’s allegation that his lawyer told him not to tell the government about Darling is not contradicted by the record, inherently incredible, conclusory, or unremediable. Moreover, if true, it might well be sufficient to satisfy the two-pronged test for establishing an actual conflict of interest: (1) that counsel “actively represented conflicting interests”; and (2) that “an actual conflict of interest adversely affected his lawyer’s performance.” See Cuyler, 446 U.S. at 349-50, 100 S.Ct. at 1719.5 The majority speculates that “Carey’s best interests could have dictated precisely the course suggested by Perrino.” See opinion at 1100 (emphasis supplied). Perhaps, but we cannot know without an evidentiary hearing at which Per-rino could either deny the allegation or explain his action. After all, the inference that Carey would have us draw from his petition — that Perrino’s advice was motivated by loyalty to Darling and could well have harmed him at sentencing — is just as plausible as the majority’s speculation. I think the law required the district court to hold a hearing and to resolve these questions.
III.
After concluding that Carey does not state an actual conflict of counsel claim, the major*1103ity states that it “still must determine whether prejudice in fact resulted.” See opinion at 1101. It does not need to engage, and should not have engaged, in this analysis. If, on the one hand, Carey’s allegation is insufficient to state an actual conflict claim (as the majority holds), there is no viable claim of ineffective assistance of counsel, and the prejudice analysis is entirely superfluous. If, on the other, Carey’s allegation is sufficient to state an actual conflict claim (as I maintain), the prejudice analysis is patently improper. The majority in Cuyler could not have been clearer: “[A] defendant who shows that a conflict of interest actually affected the adequacy of his representation need not demonstrate prejudice in order to obtain relief.” 446 U.S. at 349-50, 100 S.Ct. at 1719.
IV.
Even were I to ignore Cuyler and deem Carey’s claim as being properly subject to a prejudice analysis, I could not agree that there was no prejudice here. Hoisting Carey by the petard of an assertion Cuyler did not require him to make, the majority confines its prejudice inquiry to the harm emphasized by Carey in his motion (that his failure to talk about Darling at the initial debriefing resulted in the absence of a § 5K1.1 motion at his sentencing) and determines “with near certainty that, even if Carey had discussed Darling at the initial debriefing, any information about Darling would not have changed the government’s ultimate determination that Carey’s assistance was ‘not significant.’ ” See opinion at 1101.
I have two problems with this determination. First, I am loath to decide what the government would or would not have done in this case without sworn testimony tested by adversarial questioning. And more importantly, even if I could conclude that Carey would not have received a § 5K1.1 motion at his initial sentencing, I cannot say that Carey was not harmed in some other way if his lawyer was disloyal. In this vein, I point out that Carey could come out of a second sentencing with a lower sentence than the one he currently is serving. Carey’s guideline range was 97-121 months, and he received a sentence of 109 months. Thus, regardless of whether there was a § 5K1.1 motion, if the district court found that Carey had received ineffective assistance of counsel in connection with his first sentencing, it would in no way be engaging in an empty exercise by setting Carey’s sentence aside and ordering a second sentencing hearing. It is not at all farfetched to assume that a lawyer completely loyal to Carey might be able to persuade the judge to sentence him at the lower end of the appropriate guideline range.
The majority concludes its prejudice analysis by stating that it does not “detect any evidence even arguably suggesting that Carey’s sentencing was either unfair or unreliable.” Id. at 1101. While I don’t disagree with this statement, I hardly find it surprising; there was no evidentiary hearing at which such evidence might have been developed. That is the main point of my dissent.
V.
Binding Supreme Court and Circuit precedent prohibited the district court from rejecting Carey’s conflict of interest claim without an evidentiary hearing. I therefore dissent from Part B of the majority opinion, which affirms the denial of Carey’s claim without a hearing.
. Strickland, of course, predated the Sentencing Guidelines. Federal sentencing proceedings today can hardly be described, in the words of Strickland, as either "informal" or as governed by a senteneer with "standardless discretion.” 466 U.S. at 686, 104 S.Ct. at 2063.
It must also be noted that despite the above-quoted dictum, we have applied the Strickland effectiveness standard at sentencing. See Carsetti v. Maine, 932 F.2d 1007, 1012-14 (1st Cir.1991).
. It should be borne in mind that Cuyler's "adversity" requirement is not tantamount to a showing of harm or prejudice; Cuyler makes very clear that a defendant victimized by a lawyer with dual loyalties need not show harm or prejudice. Cuyler, 446 U.S. at 349-50, 100 S.Ct. at 1718-19. For this reason, adversity must be construed as any action taken by a lawyer which was prompted by the lawyer's loyalty to some other party and which was not in the complaining client’s interest.