UNITED STATES of America, Appellee, v. Hernando MARÍN-ECHEVERRI, Defendant, Appellant, 846 F.3d 473

KAYATTA, Circuit Judge. The defendant, Hernando Marin-Echev-erri, pled guilty to violating 21 U.S.C. §§ 952(

a) and 963 and 18 U.S.C. § 1956(a)(l)(B)(

i) and 1956(h), by participating in a conspiracy to import one or more kilograms of heroin into the United States and to launder the proceeds of that activity. The district court sentenced Marin to 262 months in prison. Marin appeals, arguing that the government violated the plea agreement during the sentencing hearing and that he received constitutionally ineffective assistance of counsel. We affirm the sentence and dismiss the ineffective assistance of counsel claim without prejudice to its reassertion in a collateral proceeding. I.

Background This appeal does not turn on the details of the crime, so we only briefly sketch the facts, drawing from “the uncontested portions of the change-of-plea colloquy, pre-sentence report, and sentencing hearing.” United States v. Gall, 829 F.3d 64, 67 n.1 (1st Cir. 2016). We set out the procedural background at greater length because it is central to Marin’s arguments on appeal. Between August 2012 and April 2013, Marin participated in a conspiracy to import between ten and thirty kilograms of heroin into Puerto Rico.

Members of the conspiracy packed suitcases full of heroin in Colombia and transported those suitcases to couriers in Venezuela. The couriers brought those suitcases into the United States. Members of the conspiracy also sent heroin to Puerto Rico via the U.S. Postal Service. Sometimes they physically transported the proceeds from the sale of the drugs from Puerto Rico back to Venezuela and Colombia, and sometimes they sent- the proceeds via wire transfers. They disguised the nature of these transfers by sending and receiving the proceeds using the names of individuals who were not part of the conspiracy.

A grand jury returned a three-count indictment against Marin, charging conspiracy to possess one kilogram or more of heroin with intent to distribute, in violation of 21 U.S.C. §§ 841(a)(1), 841(b)(l)(A)(i), and 846 (count I), conspiracy to import one kilogram or more of heroin into the United States, in violation of 21 U.S.C. §§ 952(

a) and 963 (count II), and conspiracy to launder the proceeds of these controlled substance offenses,- in violation of 18 U.S.C. § 1956(a)(1)(B)© and 1956(h) (count III). Marin agreed to plead guilty to counts II and III and the government agreed to dismiss count I. The resulting written plea agreement contains two sections germane to this appeal.

Section 7, captioned “Sentencing Guidelines Calculations,” explains that because the sentencing court is required to consider the guidelines sentencing range, “the United States and the defendant submit the following advisory Sentencing Guidelines calculations as to COUNTS TWO and THREE of the Indictment,” directing the reader to “U.S. Sentencing Commission Worksheets A, B and D attached to the instant Plea Agreement.” Those worksheets calculate Marin’s total offense level to be thirty-four.

The worksheets report the guidelines sentencing range as 151 to 188 months because, as the plea agreement explains, “[t]he sentencing guideline calculation contained [in the worksheets] is assuming a criminal history category I but there is no stipulation as to the defendant’s criminal history category.”

Section 8, captioned “Sentence Recommendation,” states that “[t]he parties agree and recommend that the Court sentence the defendant to a term of imprisonment at the lower end of the applicable guideline range determined by the Court.” At Marin’s change-of-plea hearing, both the Assistant U.S. Attorney (AUSA) and the magistrate judge mentioned the guidelines calculation in the worksheets.

After submitting the worksheets to the court and explaining how they reached a total offense level of thirty-four, the AUSA confirmed that “[t]he recommendation in this case ... would be that the defendant will serve a term of imprisonment at the lower end of the applicable guideline range determined by the Court, depending on the defendant’s criminal history category.” The magistrate judge later made sure Marin understood that “[i]n determining your sentence, the presiding judge will consider but may not follow the guidelines calculations, those calculations contained in your agreement in those worksheets.... [T]hese guidelines are of an advisory nature ... and the presiding judge may follow or may not follow them.... ” The presentence report (PSR) filed after the change-of-plea hearing but prior to the sentencing hearing placed Marin in criminal history category III.

At the same time, it disagreed with the total offense level calculation in the worksheets submitted with the plea agreement. The PSR instead calculated the total offense level as thirty-seven. The disagreement between the worksheets and the PSR as to the total offense level derives from a difference in the calculation of the adjusted offense level for the money laundering count. Both documents concluded that the total offense level for the two crimes together would be equal to the higher of the two adjusted offense levels, reduced by three levels for acceptance of responsibility.

The parties’ worksheets calculated the adjusted offense level for the conspiracy to import count to be thirty-seven, the sum of a base offense level of thirty-four and a three-level upward adjustment for playing a managerial role in the offense under U.S.S.G. § 3Bl.l(b). They calculated the adjusted offense level for the money laundering count to be twenty-two by applying U.S.S.G. § 281.1(a)(2). The PSR, by contrast, calculated the adjusted offense level for money laundering to be forty by applying U.S.S.G. § 2Sl.l(a)(l).

Section 2Sl.l(a)(

l) produced a base offense level of thirty-four, which was enhanced by four levels for playing a leadership role in the offense, U.S.S.G. § 3Bl.l(a), and by two levels because the money laundering conviction was under 18 U.S.C. § 1956, per U.S.S.G. § 281.1(b)(2)(B). ■ At the sentencing hearing, defense counsel challenged the PSR calculation of the adjusted offense level for money laundering. She argued that the calculation in the worksheets was correct and that the role-in-the-offense adjustment should have been three levels for a managerial role rather than four levels for a leadership role.

The district court rejected these arguments and adopted the calculation in the PSR. Defense counsel did not challenge the determination that Marin was in criminal history category III, and the district court also adopted this conclusion from the PSR. A total offense level of thirty-seven and criminal history category III yielded a guidelines sentencing range of 262 to 327 months, rather than the range of 151 to 188 months reported in the parties’ worksheets. After reaching this conclusion and questioning defense counsel and the defendant further, the district court asked to hear from the government.

The AUSA began, “The United States abides by the sentencing recommendation contained in the plea agreement in that the parties agreed to recommend to the Court a sentence at the lower end of the applicable guideline range....” She then answered four questions about the facts of the case. The district court sentenced Marin to 262 months in prison on count II, which was at the low end of the guidelines sentencing range determined by the court.

If the district court had agreed that the total offense level was thirty-four, as calculated on the worksheets, but still applied criminal history category III, the ■ guidelines sentencing range would have been 188 to 235 months. U.S.S.G. ch. 5, pt. A (Sentencing Table). As mentioned above, the guidelines sentencing range reported in the worksheets for a total offense level of thirty-four and criminal history category I is 151 to 188 months. Id. II.

Discussion On appeal, Marin does not argue that the district court erred in concluding that the calculations in the PSR were correct: he was in criminal history category III and had a total offense level of thirty-seven. Instead, he argues that the government breached the plea agreement. He also argues that he received ineffective assistance of counsel during plea negotiations. The government contends that a waiver-of-appeal provision in the plea agreement bars all of Marin’s arguments on appeal other than the ineffective assistance of counsel claim.

We sidestep the waiver issue, instead disposing of this appeal on the merits. See United States v. Sánchez-Maldonado, 737 F.3d 826, 827-28 (1st Cir. 2013) (“When the resolution of the underlying appeal plainly dictates affirmance, we often have elected to avoid the murky waters surrounding the waiver’s scope and proceeded to consider the merits of the appeal on the arguendo assumption that the waiver does not apply.”). A.

Breach of Plea Agreement When interpreting a plea agreement, “we construe the terms and conditions ... in accordance with traditional principles of contract law, looking outside the document only as necessary to provide illuminating context or resolve ambiguities in the writing.” United States v. Marchena-Silvestre, 802 F.3d 196, 202 (1st Cir. 2015) (citations omitted).

At the same time, in assessing compliance with a plea agreement, we frown on technical compliance that undercuts the substance of the deal. “A defendant who enters a plea agreement waives a panoply of constitutional rights and, therefore, we hold prosecutors to ‘the most meticulous standards of both promise and performance.’” United States v. Almonte-Nuñez, 771 F.3d 84, 89 (1st Cir. 2014) (quoting United States v. Riggs, 287 F.3d 221, 224 (1st Cir. 2002)). “Such standards require more than lip service to, or technical compliance with, the terms of a plea agreement.” Id.

For that reason, “it is possible for a prosecutor to undercut a plea agreement while paying lip service to its covenants.” Id. at 90-91. “We consider the totality of the circumstances in determining whether a prosecutor engaged in impermissible tactics.” Id. at 91. Marin ambitiously argues that the plea agreement bound the government to advocate for a sentence within the guidelines sentencing range calculated on the worksheets attached to the agreement.

This argument gets nowhere because the agreement.expressly makes clear that the worksheet calculations simply “assum[e] a criminal history category I but there is no stipulation as to the defendant’s criminal history category.” Thus, the agreed recommendation eschewed reference to the range reflected on the worksheet, and used as its object “the lower end of the applicable guideline range determined by the Court.” Marin drops back to arguing that the agreement at least bound the government to recommending a sentence within a guidelines sentencing range based on the total offense level used in the worksheets, allowing movement upward only for changes in calculating the criminal history category.

Marin did not raise this , argument in the district court, and so we review for plain error. See id. at 89. Like a judo move, this argument accepts the force of the government’s point that the agreement states that there is no stipulation as to criminal history category and flips it into a negative inference that there was a stipulation as to the offense level.

In Marchena-Silvestre, we were impressed with such an argument in the context of a plea agreement that confirmed a commitment to recommend a sentence within the “applicable guidelines range,” observing that the “applicable guidelines range” seemed to refer to the array of alternative ranges that were set forth in the agreement and that varied based only on alternative possibilities for the criminal history category. 802 F.3d at 198, 203.

Here though, the plea agreement did not state that the government would recommend a sentence within the “applicable guidelines range,” but rather within the “applicable guideline range determined by the Court.” So while one could well argue that an uncounseled defendant might assume that he was exposed to an upward movement' in the recommendation based only on an upward movement in his criminal history category, we cannot find—especially on plain error review—that a fair reading of the agreement plainly binds the government to such an

interpretation. See Marchena-Sil-vestre, 802 F.3d at 204 (“[T]he second prong of plain-error review ... will often have some ‘bite’ in plea-agreement cases. Not all breaches will be clear or obvious. Plea agreements are not always models of draftsmanship, so the scope of the Government’s commitments will on occasion be open to doubt.” (alteration in original) (quoting Puckett v. United States, 556 U.S. 129, 142, 129 S.Ct. 1423, 173 L.Ed.2d 266 (2009))).

Finally, Marin’s argument that the AUSA breached the plea agreement by answering the district court’s factual questions about the crime or by not challenging the admittedly correct guidelines calculations tendered in the PSR is foreclosed by Almonte-Nuñez.

As we said there: We repeatedly have emphasized that prosecutors have a ... solemn obligation to provide relevant information to the sentencing court and that a plea agreement may not abridge that obligation— [T]here is a material difference between answering questions asked by a sentencing court or bringing facts to the court’s attention and affirmatively supporting an adjustment. 771 F.3d at 90 (citations omitted); see also United States v.

Canada, 960 F.2d 263, 270 n.7 (1st Cir. 1992) (“It is necessary at all times that the government ‘level’ with the court as to the correct facts and calculations relevant to guideline sentencing.”). B. Ineffective Assistance of Counsel Marin also contends that he received ineffective assistance of counsel during the negotiation of the plea agreement and at the sentencing hearing, in violation of the Sixth Amendment. See Strickland v. Washington, 466 U.S. 668, 686, 104 S.Ct. 2052, 80 L.Ed.2d 674 (1984); see also Hill v.

Lockhart, 474 U.S. 52, 58, 106 S.Ct. 366, 88 L.Ed.2d 203 (1985) (applying Strickland to claim of ineffective assistance of counsel during plea negotiation). “Under Strickland, we first determine whether counsel’s representation ‘fell below an objective standard of reasonableness.’ We then inquire whether ‘there is a reasonable probability that, but for counsel’s unprofessional errors, the result of the proceeding would have been different,’ ” Parsley v.

United States, 604 F.3d 667, 671 (1st Cir. 2010) (citations omitted). “With regard to plea agreements, counsel has a critical obligation ...-. to advise the client of the advantages and disadvantages of a plea agreement.” Id. (alteration in original) (citations omitted). Marin argues that he received ineffective assistance of counsel because his attorney negotiated his plea agreement, and made recommendations as to whether he should accept or reject that agreement, without understanding the relevant sentencing guidelines.

This misunderstanding persisted, he claims, even when she was objecting to the PSR and arguing at the sentencing hearing. In Marin’s view, the arguments that his attorney advanced in the objection to the PSR and at sentencing make clear that she did not understand the guidelines. Her arguments assumed that the difference in total offense level between the worksheets and the PSR arose from the application of the offense grouping guideline and the two-level enhancement under U.S.S.G. § 2Sl.l(b)(2)(B).

Neither the objection nor the argument at the sentencing hearing so much as mentioned the distinction between U.S.S.G. § 2Sl.l(a)(

l) and (a)(2). Even after the district court explained at the sentencing hearing that the PSR. calculated the adjusted offense level for money laundering using U.S.S.G. § 2Sl.l(a)(l), defense counsel continued to press an argument based on grouping. Nevertheless, we decline Marin’s request that we evaluate his ineffective assistance claim on this appeal. “As a general rule, this court does not review ineffective assistance of counsel claims on direct appeal.” United States v. Vázquez-Larrauri, 778 F.3d-276, 293 (1st Cir. 2015).

This general rule results from.the fact that such claims are usually not raised in the original district court proceedings in which the defendant is represented by the lawyer said to be ineffective, and therefore the record is usually insufficient for meaningful review. Id. at 293-94. We see no reason to make an exception here. Marin has not yet raised the claim in the district court. Nor is this otherwise a situation in which “the critical facts are not genuinely in dispute and the record is sufficiently developed to allow reasoned consideration of an ineffective assistance claim.” United States v.

Natanel, 938 F.2d 302, 309 (1st Cir. 1991). To the contrary, “why counsel acted as [s]he did” is undeveloped, Vázquez-Larrauri, 778 F.3d at 294 (citation omitted), and the record does not rule out the possibility that defense counsel noticed the potential problem in the calculation of the total offense level for the money laundering count during plea negotiations and chose not to bring it to the government’s attention in the hope that it would go unnoticed, to the benefit of her client.

Thus, we dismiss the ineffective assistance .claim without prejudice to its reassertion, if the defendant so chooses, in a collateral proceeding under 28 U.S.C. § 2255. See United States v. Cardoza, 790 F.3d 247, 248 (1st Cir. 2015) (per curiam). We do think that the form of plea agreement used in this case created a nontrivial risk that the defendant would misread it unless well counseled.

A pleading defendant is usually most interested in two things concerning the meaning of a plea agreement under Federal Rule of Criminal Procedure 11(c)(1)(B): the length of the sentences in the guidelines sentencing range and the recommendation that the government will make. While we agree, as explained above, that a lawyerly reading of the agreement reveals that all it really says is that the government will recommend the low end of the range as determined by the court, whatever that may be, a lay person could easily look at this plea agreement and assume that it says something more.

After all, why bother with the worksheets otherwise—especially since they grossly underestimated the range, provided no examples of the higher ranges possible, and might have been read as implying that criminal history was the only variable? With such an agreement, the government may well risk its ability to sustain the voluntariness of the plea should the evidence support a claim that defense counsel did not explain the bait-and-switch potential. III. Conclusion For the forgoing reasons, we affirm Marin’s sentence and dismiss his ineffective assistance of counsel claim without prejudice. .

In 21 U.S.C. § 952, an amount of “heroin” means that amount of "a mixture or substance containing a detectable amount of heroin.” 21 U.S.C. § 960(a)(1), (b)(1)(A). .

Section 9 of the plea agreement repeats that the parties have not agreed to a stipulation as to Marin's criminal history category. . The PSR reached this conclusion by grouping the two counts together under U.S.S.G. § 3D1.2(

c) and applying U.S.S.G. § 3D1.3(a). The worksheets reached this conclusion by treating each count as a separate group but assigning zero units to count III under U.S.S.G. § 3D1.4(

c) based on their calculation of the offense level for that count. . The district court appears to have named the wrong count when imposing the sentence. The guidelines range of 262 to 327 months was for the money laundering count, count III. The defendant does not raise this issue on appeal. .

Section 10 of the plea agreement reports that "[t]he defendant knowingly and voluntarily waives the right to appeal the judgment and sentence in this case, provided that the defendant is sentenced in accordance with the terms and conditions set forth in the Sentence Recommendation provisions in this Plea Agreement.”