Ayestas v. Harris County: New Claims Filed During Rule 59(e) Still “Successive” Unless Within the Judgment; Nonparty Prosecutorial Work-Product Discovery Orders Are Collaterally Appealable Without Rule 72(a) Objections

1. Introduction

Parties. Carlos Manuel Ayestas (Petitioner–Appellee), a Texas capital defendant sentenced to death, litigated federal habeas claims against the State’s prosecution apparatus; the Harris County District Attorney’s Office (“HCDA”) appeared on appeal as Defendant–Appellant in its capacity as a subpoenaed third party resisting discovery.

Background. After Ayestas’s 1995 capital murder prosecution, an internal Capital Murder Summary (the “Siegler Memorandum”)—signed by then-District Attorney John B. Holmes, Jr. and drafted by Kelly Siegler—recommended seeking the death penalty partly because “the Defendant is not a citizen.” The memorandum was treated as confidential work product and not produced at trial.

Procedural posture and key issues. Ayestas’s federal habeas petition (ineffective assistance of counsel (“IAC”)) was denied in 2014. While a Federal Rule of Civil Procedure 59(e) motion was pending, habeas counsel discovered the Siegler Memorandum in a prosecution file that had been provided. Ayestas moved to amend to add Eighth and Fourteenth Amendment selective-prosecution claims. The district court treated that effort as a barred “second or successive” application under 28 U.S.C. § 2244(b). Years later, relying on Banister v. Davis, 590 U.S. 504 (2020), Ayestas filed a Rule 60(b) motion; the district court reopened the case, permitted amended pleadings, and ordered broad discovery from HCDA over privilege claims. HCDA appealed the discovery order.

The Fifth Circuit confronted (i) appellate jurisdiction to hear a nonparty’s immediate appeal of a discovery order; (ii) whether HCDA forfeited review by not objecting under Rule 72(a); and (iii) whether the district court had jurisdiction to entertain Ayestas’s selective-prosecution claims in light of § 2244(b) and the limits of Banister and Gonzalez v. Crosby, 545 U.S. 524 (2005).

2. Summary of the Opinion

The court held:

1. The discovery order is immediately appealable under the collateral-order doctrine because it compels disclosure of alleged prosecutorial work product, implicating substantial public interests recognized in selective-prosecution discovery jurisprudence.
2. Rule 72(a)’s objection requirement does not apply to nonparties by its text; HCDA’s failure to object in the district court did not defeat appellate jurisdiction.
3. The district court lacked jurisdiction over Ayestas’s selective-prosecution claims because his attempt to inject them (a) via the 2015 motion to amend while a Rule 59(e) motion was pending, and (b) via the 2021 Rule 60(b) motion, constituted barred “second or successive” claims under 28 U.S.C. § 2244(b)(2).

Because the selective-prosecution claims were jurisdictionally barred, the Fifth Circuit vacated the discovery order and dismissed the selective-prosecution claims.

3. Analysis

3.1 Precedents Cited

A. Collateral-order appealability and nonparty discovery

• La Union Del Pueblo Entero v. Abbott, 68 F.4th 228 (5th Cir. 2023), Vantage Health Plan, Inc. v. Willis-Knighton Med. Ctr., 913 F.3d 443 (5th Cir. 2019), and Lewis v. Crochet, 105 F.4th 272 (5th Cir. 2024) supplied the Fifth Circuit’s modern articulation of the collateral-order test (conclusiveness; separateness; effective unreviewability).
• Mohawk Industries, Inc. v. Carpenter, 558 U.S. 100 (2009) limited collateral-order appeals for parties asserting attorney-client privilege. The panel used Mohawk as a contrast: while parties can often obtain adequate review through post-judgment mechanisms, nonparties face distinct harms from compelled disclosure and lack control over litigation and appeal.
• Whole Woman's Health v. Smith, 896 F.3d 362 (5th Cir. 2018), Vantage Health Plan, Inc. v. Willis-Knighton Med. Ctr., and Leonard v. Martin, 38 F.4th 481 (5th Cir. 2022) were invoked to emphasize that nonparty privilege/disclosure orders can qualify for immediate review when delay would imperil “a substantial public interest” or “some particular value of a high order.”
• United States v. Nobles, 422 U.S. 225 (1975) and Hickman v. Taylor, 329 U.S. 495 (1947) grounded the work-product doctrine in systemic public policy protecting the adversarial process.
• United States v. Armstrong, 517 U.S. 456 (1996) (quoting Wayte v. United States, 470 U.S. 598 (1985)) supplied the “chilling” and “effectiveness” rationales for limiting inquiry into prosecutorial decisionmaking—rationales the panel treated as demonstrating the public-interest weight supporting immediate review.
  Notably, Armstrong’s substantive discovery bar was not ultimately applied because the panel resolved the case on jurisdictional grounds; Armstrong was used primarily to establish the importance and sensitivity of the interests at stake.

B. Successive habeas limits, Rule 59(e), and Rule 60(b)

• Banister v. Davis, 590 U.S. 504 (2020) held that a proper Rule 59(e) motion is not a “second or successive” habeas petition. The Fifth Circuit relied on Banister’s narrowing language: Rule 59(e) is backward-looking, confined to reconsideration of matters “properly encompassed” in the judgment, and cannot be used for arguments that could have been raised earlier.
• Rivers v. Guerrero, 605 U.S. 443 (2025) reinforced the conceptual “merger” of Rule 59(e) proceedings into the final judgment and explained why Rule 59(e) motions themselves are not successive filings—again underscoring that the carve-out is for the Rule 59(e) motion as such, not for any new claims filed during that window.
• Gonzalez v. Crosby, 545 U.S. 524 (2005) controlled the Rule 60(b) analysis: a Rule 60(b) motion that “seeks to add a new ground for relief” is treated as presenting a “claim” subject to AEDPA’s successive-petition gatekeeping.
• Blackman v. Davis, 909 F.3d 772 (5th Cir. 2018) was cited for the proposition that failure to satisfy § 2244(b)(2) requires dismissal for lack of jurisdiction.
• Prior Ayestas-related decisions framed diligence and procedural history, including Ayestas v. Stephens, 817 F.3d 888 (5th Cir. 2016) (quoted on lack of diligence), Ayestas v. Davis, 933 F.3d 384 (5th Cir. 2019), and Ayestas v. Thaler, 569 U.S. 1015 (2013).

C. Rule 72(a), nonparty status, and statutory-interpretation tools

• The majority’s textual reading of Fed. R. Civ. P. 72(a) (speaking only to “a party”) was supported by contrast to other rules explicitly referencing “nonparties,” and by the interpretive canon illustrated with Duncan v. Walker, 533 U.S. 167 (2001).
• On timing concerns, the majority analogized to gap-filling principles discussed in Runyon v. McCrary, 427 U.S. 160 (1976), rejecting the idea that absence of express nonparty deadlines in Rule 72(a) implies unlimited time to seek review.
• The dissent relied on magistrate-appeal limitations expressed in Mayfield v. Butler Snow, L.L.P., 75 F.4th 494 (5th Cir. 2023) and Gregg v. Linder, 349 F.3d 860 (5th Cir. 2003), and invoked the Supreme Court’s modern narrowing of collateral-order doctrine, including Will v. Hallock, 546 U.S. 345 (2006), Van Cauwenberghe v. Biard, 486 U.S. 517 (1988), and Geo Grp., Inc. v. Menocal, No. 24-758, 2026 WL 513536 (U.S. Feb. 25, 2026).
  This sets up a meaningful intra-case debate: the majority treats the order as a collateral order notwithstanding its issuance by a magistrate judge; the dissent treats direct appeal from a magistrate’s discovery order as jurisdictionally improper absent district-court review.

3.2 Legal Reasoning

A. Why the Fifth Circuit could hear the appeal

The panel treated the discovery order as an appealable collateral order because (1) it conclusively required HCDA to produce sweeping categories of prosecutorial materials, (2) the privilege/public-interest questions were separate from the merits of Ayestas’s underlying habeas claims, and (3) postponing review would cause irreparable harm by forcing disclosure of confidential prosecutorial work product—harm not remedied by later appeal because “the cat is out of the bag.”

While Mohawk Industries, Inc. v. Carpenter disfavored collateral-order appeals for privilege disputes among parties, the majority emphasized the different posture of a nonparty asserting privilege. Drawing on Whole Woman's Health v. Smith and Vantage Health Plan, Inc. v. Willis-Knighton Med. Ctr., it reasoned that third parties lack control over litigation strategy and post-judgment appeals, making immediate review more justified.

The panel also anchored “high order” interests in the prosecutorial context: work product shields the “orderly prosecution” of legal claims (United States v. Nobles; Hickman v. Taylor), and selective-prosecution discovery can chill enforcement and expose internal deliberations (United States v. Armstrong; Wayte v. United States). Those rationales supplied the “substantial public interest” needed to satisfy the “effectively unreviewable” prong.

B. Why Rule 72(a) did not block review

Ayestas argued HCDA waived appellate review by failing to object to the magistrate judge under Rule 72(a). The majority rejected that argument on textual grounds: Rule 72(a) regulates “a party,” and does not mention nonparties—unlike other federal rules that expressly do. The court treated that omission as intentional and declined to read nonparties into the rule.

The dissent sharply disagreed, contending that “party” in Rule 72 should be understood as a party to the magistrate proceeding (including a subpoenaed nonparty), and warning that the majority’s approach creates a new, unnecessary bypass of district-court review of magistrate orders.

C. Why the district court lacked jurisdiction over the selective-prosecution claims

The jurisdictional core is AEDPA’s successive-petition bar. Under 28 U.S.C. § 2244(b)(2), new claims in a “second or successive” habeas application must fit narrow exceptions (new retroactive constitutional rule, or newly discovered facts plus a stringent innocence-related showing). If not, courts must dismiss for lack of jurisdiction (Blackman v. Davis).

1) The 2015 Motion to Amend was successive despite being filed during a pending Rule 59(e) motion.
Ayestas relied on Banister v. Davis to argue that the pendency of Rule 59(e) kept the habeas proceeding “open” such that amendment could not be “successive.” The panel distinguished Banister by focusing on its limits: Rule 59(e) permits reconsideration of matters encompassed in the judgment and bars raising arguments that could have been made earlier. Ayestas’s new selective-prosecution theories (Eighth/Fourteenth) were not a reconsideration of the IAC judgment; they were different claims dependent on a different factual predicate (the Siegler Memorandum).

The panel additionally rejected the “suspends finality” argument. Relying on Rivers v. Guerrero, it held that the “merger” concept protects Rule 59(e) motions themselves from being labeled successive; it does not transform any filing made during the 59(e) window into a non-successive pleading. The court warned that accepting Ayestas’s theory would allow petitioners to file placeholder Rule 59(e) motions and then “flood” courts with new claims, effectively defeating Rule 59(e)’s narrow function.

Finally, the panel treated the factual predicate as discoverable earlier with due diligence, quoting Ayestas v. Stephens for the absence of evidence that counsel tried to examine the prosecution file before the search that found the memorandum. Because the claim neither relied on a new retroactive constitutional rule nor satisfied the due-diligence exception in § 2244(b)(2)(B), it was barred.

2) The 2021 Rule 60(b) motion was an improper vehicle for adding new claims.
Invoking Gonzalez v. Crosby, the panel held that using Rule 60(b) to add new grounds for relief circumvents AEDPA. Even though Ayestas framed the motion as seeking relief from a procedural ruling (the earlier denial of amendment), the substance was to present selective-prosecution “claims” not included in the original habeas petition. Because those claims were successive under § 2244(b)(2) and did not qualify for either exception, the district court lacked jurisdiction to entertain them or to order discovery in aid of them.

3.3 Impact

A. Habeas practice: narrowing the practical reach of Banister

The decision draws a bright operational line: Banister v. Davis protects a proper Rule 59(e) motion as a continuation of the first habeas proceeding, but it does not authorize using the 59(e) window as a staging ground for brand-new constitutional claims. In the Fifth Circuit, a petitioner who discovers new facts during the 59(e) period cannot assume amendment is “non-successive”; the new claim must still be “properly encompassed” within reconsideration of the just-issued judgment, or it risks being treated as a barred successive application.

B. Rule 60(b): reaffirming Gonzalez’s substance-over-labeling approach

The opinion reinforces that intervening procedural decisions about how habeas rules interact (here, Banister’s reading of Rule 59(e) vis-à-vis § 2244(b)) do not open the door to Rule 60(b) claim expansion. The Fifth Circuit’s approach encourages strict separation between (i) “true” procedural attacks that reopen merits consideration of already-presented claims and (ii) attempts to introduce new constitutional theories.

C. Immediate appeals by subpoenaed prosecutors and privilege holders

On appealability, the majority’s combination of (1) collateral-order availability for a nonparty and (2) a textual exclusion of nonparties from Rule 72(a)’s objection requirement creates a potentially significant pathway for third-party governmental entities to obtain immediate circuit review of magistrate-issued discovery orders that implicate prosecutorial work product. The dissent forecasts systemic consequences: direct appellate review of magistrate discovery orders could become more common, potentially weakening district judges’ supervisory role and stressing finality norms.

D. Selective-prosecution claims in habeas: practical barriers

Although the court did not reach the merits of selective prosecution or the discovery standard under United States v. Armstrong, the jurisdictional holding functions as a powerful gatekeeping tool: selective-prosecution theories discovered late must clear AEDPA’s successive-petition barriers, including due diligence. That is a high practical hurdle because selective-prosecution evidence often resides in prosecutorial files, and diligence disputes may be decisive.

4. Complex Concepts Simplified

“Second or successive” habeas application (28 U.S.C. § 2244(b))

AEDPA generally allows one full federal habeas challenge. Later efforts to add new claims are presumptively barred unless they fit narrow exceptions. If the bar applies, the federal court lacks power (jurisdiction) to hear the claim.

Rule 59(e) vs. Rule 60(b)

Rule 59(e) is a short, 28-day mechanism for asking the court to reconsider issues already decided in the judgment—typically to correct errors. Banister v. Davis holds such motions are part of the first habeas case, not a second petition. Rule 60(b) is a later, limited remedy for extraordinary circumstances (mistake, fraud, etc.), but Gonzalez v. Crosby forbids using it to introduce new habeas claims without satisfying AEDPA’s successive-claim rules.

Collateral-order doctrine

Most appeals must wait until the case ends. A small class of orders can be appealed immediately if they conclusively decide an important, separate issue that would be effectively unreviewable later. The Fifth Circuit treated compelled disclosure of a nonparty prosecutor’s work product as fitting that narrow class.

Work-product doctrine (prosecutorial work product)

Protects an attorney’s (or prosecutor’s) internal materials reflecting strategy, mental impressions, and legal theories. The doctrine aims to preserve candid preparation and prevent forced disclosure of deliberative processes.

Selective prosecution

A constitutional claim alleging the government singled out a defendant for harsher treatment based on an impermissible factor (e.g., race, nationality). Courts impose a demanding threshold for discovery because probing charging decisions risks chilling legitimate law enforcement (United States v. Armstrong).

5. Conclusion

Ayestas v. Harris County establishes two practical rules with wide procedural reach in the Fifth Circuit: (1) Banister v. Davis does not allow petitioners to use a pending Rule 59(e) motion as a vehicle to add new habeas claims unrelated to reconsideration of the judgment; such claims remain “second or successive” and must satisfy 28 U.S.C. § 2244(b)(2). (2) A nonparty prosecutor asserting work-product concerns may obtain immediate appellate review of a compelled-disclosure order as a collateral order, and the panel’s textual approach holds that Rule 72(a)’s objection requirement does not bind nonparties.

The immediate consequence was decisive: lacking jurisdiction over the selective-prosecution claims, the district court could not compel discovery to support them; the Fifth Circuit therefore vacated the discovery order and dismissed the claims. More broadly, the decision tightens AEDPA gatekeeping around late-discovered prosecutorial-motive theories while expanding (at least in this setting) a nonparty’s ability to seek prompt appellate protection against compelled disclosure of prosecutorial files—an expansion the dissent warns may erode the traditional hierarchy of magistrate and district judge review.