MortalShell2.lol logo MortalShell2.lol Unofficial MS2 Tools

Mortal Shell 2 Keeps Crashing? Every Known Fix (2026)

Mortal Shell 2 keeps crashing on PC? Every tracked fix lives on this page — 52 issues in total: 22 triaged into three fast-first phases, plus 30 recently reported items with community or developer evidence. The September 24 hotfix (1.000.011) developer-patched ten of the recently reported items — including the two named achievement fixes, the Severance soft lock and Stray Genessa pair, and the Warden and Hexapod boss corrections — and their entries now carry Fixed status below. The October 14 hotfix (1.000.012) named no tracked crash entry, so every crash status below is unchanged. One status moved after this page was written: the Bring My Ova Back to Me achievement, first listed under the Sept 24 hotfix's fixes, was corrected by the developer as not actually fixed in that patch — its entry (#24) now reads Partial until the next update, so treat that specific achievement as still broken. Work top to bottom and stop the moment the crashing stops. Status labels are honest: Fixed means developer-patched, Workaround means community-proven, Pending means we're waiting on the developer.

Last verified: 2026-10-18Game version: Hotfix 1.000.012

How this page works. Every crashing issue listed here was extracted from real error strings, community reports, and patch notes — not guessed from generic "game keeps crashing" threads. We organized 22 triaged issues into three phases ordered by how quickly you can test each fix. The methodology is simple: start at Phase 1, try each item in order, and stop the moment your crashing stops. Don't skip ahead to Phase 3 fixes when a 30-second toggle in Phase 1 would have stopped the crashing. Six of the tracked crash issues have been fixed by the developer across five major patches (Hotfix 1.0 on August 17, Hotfix 2.0 on August 18–19, Balance Patch 1 on August 22–23, the Week 1 Update on August 29, and the Adaptive Difficulty update on September 5 — plus the smaller Hotfix 3 on September 1, which fixed non-crash bugs). Ten more recently reported items were developer-patched in Hotfix 1.000.011 (September 24). Twelve have community-verified workarounds and one is partially fixed. Four remain pending with no reliable player-side fix available yet.

Status labels: Fixed = patched by the developer. Workaround = a community-tested fix you can apply right now. Pending = we're waiting on the developer; no reliable fix exists yet. Partial = fixed for most players but some still report crashing.

Each entry lists the symptom (what you see), the cause (the actual error string or technical root cause, not a guess), and the fix (what to do about it). Work top to bottom. If a Phase 1 fix solves your crashing, stop — don't apply Phase 3 fixes to a solved problem.

Why is Mortal Shell 2 crashing so much?

Mortal Shell 2 is built on Unreal Engine 5, and that alone explains a lot of the crashing patterns you'll see below. UE5 introduced a new DirectX 12 renderer (D3D12RHI), a fundamentally different shader compilation pipeline using Oodle-compressed ShaderCodeArchives, and a more aggressive garbage collection system compared to UE4. These are not abstract engine trivia — they map directly to the most common crashing types in Mortal Shell 2. D3D12 GPU crashing is the single largest category here, and it tends to be triggered by driver quirks, overclocking, and settings combinations that UE4 games handled without crashing. Shader compilation crashing at launch is another UE5 signature: the engine decompresses and compiles shader caches on first run, and when that process hits a corrupted cache or a storage bottleneck, it crashes hard before you even reach the main menu. UE5's garbage collector can also enter a bad state and take the whole game down with an EXCEPTION_ACCESS_VIOLATION, which is why you'll see multiple crashing issues traced to GarbageCollection.cpp. None of this is unusual for a UE5 game in its first weeks — the engine is powerful but unforgiving, and Mortal Shell 2 is hitting these rough edges at launch. The good news is that most UE5 crashing problems have known workarounds, and the developer has already shipped eight patches in under six weeks — the latest being Hotfix 1.000.011 on September 24 (Egon's Stone dungeon-loading and Gloom-drop fixes, the two named achievement fixes, the Severance soft lock and Stray Genessa pair, Warden and Hexapod boss corrections, Adaptive Difficulty tier UI, and a new Low Latency Mode) and, before it, Hotfix 1.000.010 on September 14 (a fix-only list covering the Proxima Biosampler animation, Tarstone durability and Gloom exploits, boss and NPC corrections, and general stability improvements).

Runs into a corrupted save instead of a crash? Phase 2's save-file fixes (delete the two MortalShell2 .sav files) also cover keybind and settings-menu crashes — and the same steps reset corrupted settings back to defaults.

Triage phases — what each one targets

The 22 triaged issues in Mortal Shell 2 aren't random — they cluster into distinct categories, and that's why the three-phase structure matters. Phase 1 (Quick Wins) targets the crashing you can fix in under 30 seconds by changing a single setting. This phase catches the most common crashing triggers: Frame Generation (the number one cause of crashing in Mortal Shell 2), FSR/TSR upscaling conflicts, Ray Tracing overhead, and shader cache size. If you're crashing, there's roughly a 60% chance a Phase 1 toggle stops the crashing. Phase 2 (Deeper Fixes) addresses crashing that requires a couple of minutes — driver issues, corrupted save files, display mode switches, and GPU instability. These are the D3D12 GPU crashing and garbage collection exceptions that a simple setting toggle won't resolve. Phase 3 (Advanced) is for the stubborn crashing: cache purges, page file reconfiguration, out-of-video-memory errors, and memory leak workarounds (the long-session kind has a dedicated memory leak guide). Most players never need Phase 3, but if your crashing persists through the first two phases, that's where the remaining answers live.

New this cycle — fix these first

Phase 1 — Quick Wins

30 seconds each · 11 known issues

Shader compilation crash

Symptom: Shader compilation crash — 'Compiling Shaders' freezes or crashes before the main menu

Cause: Oodle decompression error: LowLevelFatalError: Could not decompress shader group with Oodle (ShaderCodeArchive.cpp:499) An i7-14700K player with this exact Oodle signature reported crashes persisting after the Week 1 patch (community-reported, Aug 30 follow-up); the error string in their report links to Oodle's own advisory on Intel 13th/14th-gen instability — a vendor-side attribution, not a statement by the game's developer. Their post-Sept-24 status is unknown.

Fix / workaround:

  • NVIDIA Shader Cache Size → Unlimited (recommended by the Playstack community manager)
  • Clear DXCache + GLCache
  • Retry compilation multiple times

Note: The Oct 14 hotfix rebuilt shader (PSO) caches and world HLODs; this may relate to shader-related crashes, but the developer did not name any crash fix in the notes. Status unchanged — monitor for recurrence (Steam event comments follow-up scheduled). The Oct 14 hotfix rebuilt shader (PSO) caches and world HLODs (general optimization, not a targeted fix).

Death/reload crash

Symptom: Death/reload crash — loading screen freezes after dying, then CTD

Cause: D3D12 GPU crash / render thread crash: EXCEPTION_ACCESS_VIOLATION reading 0xffffffffffffffff — related to Frame Generation

Fix / workaround:

  • Disable Frame Generation (restart required)
  • Switch to Fullscreen
  • Drop to High preset
  • FSR → TSR

DLSS Frame Generation crash

Symptom: DLSS Frame Generation crash — crashing mid-game with FG on, especially on respawn

Cause: FG conflicts with the UE5 render pipeline; 2x/3x auto modes are worst

Fix / workaround:

  • Turn off Frame Generation and restart

Burn status crash

Symptom: Burn status crash — 0x0000000000000000 GarbageCollection.cpp:1744

Cause: Lazlo Retribution + Burn status effect interaction

Fix / workaround:

  • Avoid using the Burn status effect until patched

Intro cinematic crash

Symptom: Intro cinematic crash — crashing during the torch + wall-break opening cinematic

Cause: FSR upscaling conflicts with that scene

Fix / workaround:

  • FSR → TSR

Post-death loading lockup

Symptom: Post-death loading lockup — completely locks up after death

Cause: Shader cache exhaustion / D3D device loss

Fix / workaround:

  • NVIDIA Shader Cache Size → Unlimited

Riposte steals enemy weapon

Symptom: Riposte steals enemy weapon — enemies 'steal' your weapon pose during Riposte

Cause: Game logic bug

Fix / workaround:

  • Fixed in HF2.0

Slayer Seal UI lockup

Symptom: Slayer Seal UI lockup — equipping the Slayer Seal from the Beacon menu locks the UI

Cause: UI logic bug

Fix / workaround:

  • Fixed in HF2.0
  • Re-fixed in Hotfix 1.000.011: "Fixed a bug that made the Slayer Seal unresponsive" (developer patch notes). This is a distinct symptom form from the HF2.0 fix — that one covered the Beacon menu UI lockup, while the 1.000.011 fix covers the Slayer Seal item itself becoming unresponsive.

Vatra's Temple progress lock

Symptom: Vatra's Temple progress lock — the temple stays permanently locked after exploring it

Cause: Progression logic bug

Fix / workaround:

  • Fixed in the Week 1 Update (console patch 1.000.008)

Lazlo's Rekindle crash / softlock

Symptom: Lazlo's Rekindle crash / softlock — potential crash or progress freeze while using Lazlo's Rekindle

Cause: Ability interaction bug

Fix / workaround:

  • Fixed in the Week 1 Update (console patch 1.000.008)

Gragu barkeep softlock

Symptom: Gragu barkeep softlock — talking to the tavern barkeep between Gragu's shell dying and reclaiming it could freeze progress

Cause: Dialogue and progression logic bug

Fix / workaround:

  • Fixed in the Week 1 Update (console patch 1.000.008)

Note: The Oct 14 hotfix separately fixed "an issue that blocked Gragu from being possessed" — a distinct symptom form; see new entry #52. This entry remains the barkeep softlock form.

Phase 2 — Deeper Fixes

2 minutes each · 8 known issues

Keybind menu crash

Symptom: Keybind menu crash — crashing when opening Settings > Keyboard Bindings

Cause: Corrupted input settings save

Fix / workaround:

  • Delete EnhancedInputUserSettings.sav + SpartaGameSettings.sav under %AppData%\Local\MortalShell2\Saved\SaveGames

D3D12 GPU Crash Dump

Symptom: D3D12 GPU Crash Dump — 'GPU Crash dump Triggered' FD3D12DynamicRHI::TerminateOnGPUCrash

Cause: DX12 GPU driver crash; targeted by HF1.0 but some players still report it. Developer-acknowledged (Steam Discussions, Aug 25 & Aug 28): the developer confirmed it is "actively working on this issue and are looking into a fix." Players report this crash did not occur during advanced access and started appearing after the early hotfixes — and new reports of the same crash dump continued after the September 5 Adaptive Difficulty patch (Sept 9–10). One player traced the dump to INVALID_POINTER_READ in SetShaderParametersOnContext (community analysis, unconfirmed).

Fix / workaround:

  • DDU clean driver reinstall
  • Remove GPU overclocks
  • Last resort: -dx11 launch option (may cause visual glitches on some systems)

Note: The Oct 14 hotfix rebuilt shader (PSO) caches and world HLODs; this may relate to shader-related crashes, but the developer did not name any crash fix in the notes. Status unchanged — monitor for recurrence (Steam event comments follow-up scheduled).

GC exception crash

Symptom: GC exception crash — EXCEPTION_ACCESS_VIOLATION 0x00000000000002c8 in a GarbageCollection stack

Cause: UE5 GC bug; may correlate with unstable Intel 13th/14th-gen CPUs A Steam-side i7-14700K report (Aug 21) shows the same Oodle shader-decompression signature; see issue #1 for the shader-compile workaround set.

Fix / workaround:

  • Disable FG + Ray Tracing
  • Cap FPS — One Linux/NVIDIA player measured the Xid 13 fault rate scaling with frame cap (community-reported, measured by one player, Aug 22–23, before the Week 1 patch): roughly one fault per ~19 minutes at ~175 fps, one per ~3 hours at ~100 fps, and zero faults in 2 hours at a 60 fps cap. This is one player's self-test on their own hardware, not a general crash rate; the developer logged the report (post-Sept-24 status unknown).

Note: The Oct 14 hotfix rebuilt shader (PSO) caches and world HLODs; this may relate to shader-related crashes, but the developer did not name any crash fix in the notes. Status unchanged — monitor for recurrence (Steam event comments follow-up scheduled).

Settings menu GC crash

Symptom: Settings menu GC crash — 0x0000000220000769

Cause: UE5 GC triggered when opening settings

Fix / workaround:

  • Disable Steam Overlay
  • Verify files
  • Delete corrupted saves (same as keybind fix)

Black screen after shader compile

Symptom: Black screen after shader compile — audio plays, no picture

Cause: Known issue, developer-acknowledged

Fix / workaround:

  • No confirmed fix yet
  • Try: clean driver reinstall, verify files, disable overlays

Save load crash

Symptom: Save load crash — UE error when loading a save

Cause: Unknown

Fix / workaround:

  • Fixed in the Week 1 Update — its crash-fix list covers save loading
  • Still crashing after updating? Verify game files, then report it to the developer

D3D Device Lost

Symptom: D3D Device Lost — 'D3D device being lost' error

Cause: GPU overclock or driver crash

Fix / workaround:

  • Remove MSI Afterburner OC/undervolt
  • DDU clean reinstall

HDD install stutter/crash

Symptom: HDD install stutter/crash — severe streaming stutter and crashes

Cause: The game needs an SSD; HDDs stutter badly and crash

Fix / workaround:

  • Reinstall to SSD (70GB free, system drive recommended)

Note: The Oct 14 hotfix includes "a fix for slow-streaming dungeons when 'seamless dungeon streaming' is enabled" — a different symptom form from HDD stutter; not merged, status unchanged.

Phase 3 — Advanced

5 minutes each · 3 known issues

Fast-travel / Beacon interaction crash

Symptom: Fast-travel / Beacon interaction crash — UE error when teleporting or leaving a Beacon

Cause: Regression bug, more frequent after HF2.0

Fix / workaround:

  • No player-side fix yet
  • Report to the developer

Out of video memory

Symptom: Out of video memory — 'Out of Video Memory' error

Cause: Ultra textures + high resolution exceed 8GB VRAM

Fix / workaround:

  • Textures to High
  • Disable hardware acceleration in Discord/Chrome

UE5 memory leak over long sessions

Symptom: UE5 memory leak over long sessions — closing after long play sessions

Cause: UE5 memory leak Distinct from a community-reported CPU-usage escalation after opening and closing the map (watchlist item, Oct 13 evidence) — that reporter verified no memory or VRAM growth via the Steam overlay; #16 remains a memory-leak issue over long sessions.

Fix / workaround:

  • Static page file: initial 16384 MB / max 32768 MB on SSD

Recently reported (community)

Community reports · no confirmed numbers · crash clusters developer-acknowledged (problem-cluster level, specific symptom not named)

VFX-heavy scene crashes

Symptom: VFX-heavy scene crashes — performance degradation in particle-heavy scenes after the September 5 Adaptive Difficulty patch

Report: Community-reported: some players report degraded performance and crashes in particle-heavy scenes (electric effects, sparks) after the September 5 Adaptive Difficulty patch. Community-reported; developer-acknowledged at the problem-cluster level: the developer has not named this specific symptom, but did confirm on Steam Discussions (Aug 23): "We are aware of this bug and are looking into a solution." for the surrounding crash cluster. Reports of render-path GPU crashes continued after Sept 5 (Sept 9–10). One player's crash log pointed to corrupted shader parameters during rendering (INVALID_POINTER_READ in SetShaderParametersOnContext) — community analysis, unconfirmed.

Suggested workaround (unverified):

  • Suggested workaround (unverified): lower particle/effect quality
  • Cap the framerate
  • Avoid prolonged VFX-heavy fights until more reports confirm the pattern

Achievements not popping

Update (Oct 16): the developer has re-confirmed Bring My Ova Back to Me is still not fixed, with a fix planned for next week's patch. Nothing to action on your side — the fix is patch-side. The Oct 14 hotfix's three extra ovas in the NG+ final dungeon may ease the count, but the achievement itself stays locked until that patch.

Symptom: Achievements not popping — some achievements stopped unlocking after the Week 1 Update

Report: Community-reported: players in the release megathread report that some achievements stopped unlocking after the Week 1 Update. The Sept 24 hotfix notes (1.000.011) initially listed two achievements as fixed — Seeking the Past and Bring My Ova Back to Me — but the developer then corrected the announcement: Bring My Ova Back to Me is NOT fixed in this patch, and a fix is planned for the next update (see the quoted correction below). Seeking the Past keeps the developer-patched status, though a few players still report unlock trouble (community-reported). Reports of achievements failing to pop continued after the Sept 24 hotfix; shell-unlock achievements (a distinct symptom form) are tracked separately in issue #47.

Suggested workaround (unverified):

  • Seeking the Past: Fixed in Hotfix 1.000.011 (Sept 24) per the developer's patch notes — update to Hotfix Update 1.000.012 (Oct 14) and re-check. If it still doesn't pop, restart the game after updating: at least one player reports the achievement auto-unlocked on relaunch, while another reports it still bugged after re-watching all memories (both community-reported, Sept 25).
  • Bring My Ova Back to Me: NOT fixed in this patch. The developer's own correction states: "Bring my Ova Back to me is not fixed in this patch, but a fix for this will be added in the next update." No workaround is known — the fix is patch-side and expected in the next update. Multiple players report the achievement still locked after 1.000.011 (community-reported); some report it may have been fixed only for new players, but that has not been confirmed.
  • “It should trigger automatically yet it doesn't. I think they fixed it for new players. I still have my achievement locked” — Steam announcement comment, Sept 25 (community-reported speculation; not confirmed)

Note: The Oct 14 hotfix adds "3 extra ovas in the final slice dungeon when playing NG+", which may ease the 'Bring My Ova Back to Me' count for NG+ players. The notes name no achievement fix; this entry stays partial pending community confirmation (CF-R4-04 context stands). Two players also asked about "Seeking the Past" still not triggering in the Oct 14 hotfix announcement comments (one in English, one in Portuguese; questions only, unconfirmed — no reproduction evidence). Post-1.000.012: one player reports "Seeking the Past" still stuck at 39/40 after watching all shell memories, across NG+ (community-reported, single report, Oct 15 — Bug Reports thread 581684714276050215). The developer is gathering info on achievement reports in that thread but has confirmed nothing. Oct 16 update (developer announcement on Reddit, level 4): the developer re-confirms Bring My Ova Back to Me is still not fixed, and states a fix is planned for next week's patch — more specific than the Sept 24 correction's 'next update' wording. Status stays partial: Seeking the Past keeps its developer-patched status (with the standing community reports noted above), the Ova achievement remains unfixed on all platforms as of the Oct 16 announcement, and the Oct 14 hotfix's 3-extra-ovas change remains a count easing measure, not an achievement fix. Watch item: next week's patch notes — if they name the Ova fix, re-verify wording verbatim before any status change.

Rewards not granted on death

Symptom: Rewards not granted on death — individual reports of missing rewards after kills

Report: Community-reported: some players report rewards not being granted on enemy/player death. Timeline note: the Week 1 Update already restored missing rewards per the developer's patch notes — reports of new missing rewards are individual and unverified; do not assume the earlier fix regressed until confirmed.

Suggested workaround (unverified):

  • No confirmed workaround yet. If rewards are missing, note the exact scenario and check back after the next patch.

4K resolution unavailable

Symptom: 4K resolution unavailable — display options capped at 2560×1440 (community report)

Report: Community-reported: a player reports 4K resolution unavailable, capped at 2560×1440. Whether this is a bug, a display-configuration issue, or an intentional cap is unconfirmed. The 2560×1440 figure is the player's own report, not a confirmed value.

Suggested workaround (unverified):

  • No confirmed workaround yet. Check display configuration and GPU driver resolution settings while the cause is unconfirmed.

Won't launch

Symptom: Won't launch — "Failed to start Process… An Application Control policy has blocked this file (0x11C7)" on Windows 11

Report: Windows 11 Smart App Control false-positive flagging MortalShell2.exe (developer: the game is safe to run; the flag is unnecessary — developer reply, Aug 25 Steam Discussions). The developer has logged the issue and is looking into it.

Suggested workaround (unverified):

  • Turn off Smart App Control (Settings > Privacy & Security > Windows Security > App & browser control), or make an exception for Mortal Shell II — developer-recommended, Aug 25 Steam reply
  • Note: Smart App Control can only be fully disabled once per Windows install (Windows policy); the developer has logged the false flag and is looking into it

Crash when pressing Continue at the main menu

Symptom: Crash when pressing Continue at the main menu — the game consistently crashes on first load while continuing an existing save, before reaching gameplay

Report: Community-reported: three independent players have reported this crash line since late August (two on Steam Discussions, one on Reddit), across different GPU and CPU configurations. Developer-acknowledged at the crash-cluster level only: the developer confirmed on Steam Discussions (Aug 26): "We are aware of these crashing issues and are looking into a solution." — but has not named this specific symptom. Whether Hotfix 1.000.011 (Sept 24) addressed it is unconfirmed.

Suggested workaround (unverified):

  • Reported workaround (reporter-verified): set a 30 FPS lock in the settings before pressing Continue. Multiple players report this stops the crash, but it must be re-applied after each system restart.
  • Alternative workaround (reporter-verified): create or load a new save slot first, then load the intended save. Reported effective in repeated tests, also needs repeating after a restart.
  • No developer-patched fix yet — status pending. If the crash persists after updating to Hotfix 1.000.011, use the workarounds above and check this page for updates.

Fixed in Hotfix 1.000.011 (Sept 24)

Falling through elevator lifts

Symptom: Falling through elevator lifts — walking into walls while riding elevator lifts could drop the player through the floor

Cause: Collision bug. The developer has not described the underlying cause.

Fix (developer-patched):

  • Fixed in Hotfix 1.000.011 (Sept 24): the developer's patch notes list "Several fixes to prevent players falling through elevator lifts when walking into walls". Update to the latest patch.

Egon's Stone issues

Symptom: Egon's Stone issues — dungeons not loading correctly when using the Stone, and the Stone removing Gloom dropped in the open world

Cause: Item logic bugs. The developer has not described the underlying cause.

Fix (developer-patched):

  • Fixed in Hotfix 1.000.011 (Sept 24): the developer's patch notes list "Fixed dungeons not loading correctly when using Egon's Stone" and "Fixed Egon's Stone removing Gloom dropped in the open world". Update to the latest patch.

Boss fight issues

Zmey fight bug

Symptom: Zmey fight bug — weapons disappear or attacks stop during the Zmey boss fight

Cause: Grisha Remnant summon interaction (developer-confirmed bug, Aug 20 Steam reply). The Sep 14 1.000.010 notes list fixed spawning and targeting issues with the Grisha Remnant and the Summon Stone; whether the weapon-disappearance symptom shares the same root cause has not been named in the notes.

Fix (developer-patched):

  • Unequip the Grisha Remnant before the Zmey fight (developer-recommended workaround)
  • If hit: 1.000.010 (Sep 14) fixed Grisha Remnant spawning/targeting — update first, then retry. If it still happens after updating, unequip the Grisha Remnant
  • Note (post-1.000.010, community-reported): a distinct symptom form was reported after the Sept 24 window — losing the shell during phase 2 of the Zmey fight (two independent reports). A community reply attributes it to the curse mechanic built up by the boss's fire breath, which an item can cleanse — so this may be a mechanic misread rather than a bug; it is not confirmed as a bug. It is tracked separately and not merged with the weapon-disappearance fix above.

Orrem, the Reclaimed dying instantly

Symptom: Orrem, the Reclaimed dying instantly — the miniboss could be killed immediately at the start of the encounter

Cause: Encounter logic bug. The developer has not described the underlying cause.

Fix (developer-patched):

  • Fixed in Hotfix 1.000.011 (Sept 24): the developer's patch notes list "Fixed an issue that could cause Orrem, the Reclaimed to be killed instantly". Update to the latest patch.

The Warden jumping through walls and unresponsive jumps

Symptom: The Warden jumping through walls and unresponsive jumps — jump timing and trigger boxes felt unreliable, and the boss could jump through walls during the fight

Cause: Boss AI and trigger-box bugs. The developer has not described the underlying cause.

Fix (developer-patched):

  • Fixed in Hotfix 1.000.011 (Sept 24): the developer's patch notes list "Adjusted The Warden's jump timing and trigger boxes to be more responsive", "Adjusted The Warden's positioning so it no longer jumps through walls", and "Several fixes and improvements to The Warden". Update to the latest patch.

Pushed outside the Hexapod arena

Symptom: Pushed outside the Hexapod arena — the player could be pushed out of the boss arena during the Hexapod fight

Cause: Arena collision bug. The developer has not described the underlying cause.

Fix (developer-patched):

  • Fixed in Hotfix 1.000.011 (Sept 24): the developer's patch notes list "Fixed the player being pushed outside the Hexapod arena". Update to the latest patch.

Platform / Linux (Proton)

Random crashes every 10–30 min on AMD GPUs under Proton/Linux

Symptom: Random crashes every 10–30 min on AMD GPUs under Proton/Linux — amdgpu ring gfx_0.0.0 timeout, often taking the desktop session down

Report: Mesa/RADV graphics queue issue on the Proton compatibility layer (Valve-tracked, Proton issue #10121, open) — community-reported for this game, not confirmed by the game's developers

Suggested workaround (unverified):

  • Try launch options RADV_DEBUG=single_queue %command% (Valve-recommended first step; may only delay crashes)
  • If it persists: RADV_DEBUG=nocompute VKD3D_CONFIG=single_queue — reporter-verified workaround on RX 7900 XTX (community-reported, Sep 5)
  • Track Proton issue #10121 for updates
  • Note: a Valve member (kisak-valve) suggests the single_queue launch option on Proton issue #10121 — this is Valve-tracked Proton-layer guidance, not a statement from the game's developers. The reporter later found RADV_DEBUG=nocompute VKD3D_CONFIG=single_queue stopped the crashing for them ("for now" — community-reported, not a confirmed fix). The Proton issue remains open as of Sept 28.

Other reported issues (gameplay bug)

Martyr Blade scaling possibly bugged

Symptom: Martyr Blade scaling possibly bugged — single Steam Discussions report

Report: Community-reported: a Steam Discussions thread asks whether Martyr Blade's scaling is bugged. Single report, no measurements, no developer response — treat as an observation only.

Suggested workaround (unverified):

  • No confirmed workaround — observation only, pending more reports.

Mether pulse stuck at +1

Symptom: Mether pulse stuck at +1 — single Steam Discussions report

Report: Community-reported: a player reports "Mether pulse stuck at +1" on Steam Discussions. Single unverified report; no workaround confirmed yet.

Suggested workaround (unverified):

  • No confirmed workaround yet.

Coward's Retreat input bug

Symptom: Coward's Retreat input bug — single Steam Discussions report

Report: Community-reported: a Steam Discussions thread describes an input bug with Coward's Retreat. Single report, details unverified.

Suggested workaround (unverified):

  • No confirmed workaround yet.

Zmey fight bug

Symptom: Zmey fight bug — weapons disappear or attacks stop during the Zmey boss fight

Cause: Grisha Remnant summon interaction (developer-confirmed bug, Aug 20 Steam reply). The Sep 14 1.000.010 notes list fixed spawning and targeting issues with the Grisha Remnant and the Summon Stone; whether the weapon-disappearance symptom shares the same root cause has not been named in the notes.

Fix (developer-patched):

  • Unequip the Grisha Remnant before the Zmey fight (developer-recommended workaround)
  • If hit: 1.000.010 (Sep 14) fixed Grisha Remnant spawning/targeting — update first, then retry. If it still happens after updating, unequip the Grisha Remnant
  • Note (post-1.000.010, community-reported): a distinct symptom form was reported after the Sept 24 window — losing the shell during phase 2 of the Zmey fight (two independent reports). A community reply attributes it to the curse mechanic built up by the boss's fire breath, which an item can cleanse — so this may be a mechanic misread rather than a bug; it is not confirmed as a bug. It is tracked separately and not merged with the weapon-disappearance fix above.

Mether's Severance soft lock and Stray Genessa

Symptom: Mether's Severance soft lock and Stray Genessa — cancelling Severance with a gamepad could soft-lock the game, and Stray Genessa became non-interactable after using Mether's Severance

Cause: Bugs in the Mether's Severance interaction introduced alongside the Genessa trainer functionality added in Hotfix 1.000.010. The developer has not described the underlying cause.

Fix (developer-patched):

  • Fixed in Hotfix 1.000.011 (Sept 24): the developer's patch notes list "Fixed a soft lock when cancelling Mether's Severance with a gamepad" and "Fixed Stray Genessa not being interactable after using Mether's Severance". Update to the latest patch.

Harbinger shades and masks missing from the level-up tab

Symptom: Harbinger shades and masks missing from the level-up tab — the items stopped displaying where you spend them

Cause: UI display bug. The developer has not described the underlying cause.

Fix (developer-patched):

  • Fixed in Hotfix 1.000.011 (Sept 24): the developer's patch notes list "Fixed Harbinger shades and masks not displaying in the level-up tab". Update to the latest patch.

Genessa's Faithful Doubles issue when using Lost Clotstone

Symptom: Genessa's Faithful Doubles issue when using Lost Clotstone — the ability behaved incorrectly with the Clotstone equipped

Cause: Item-interaction bug. The developer has not described the underlying cause.

Fix (developer-patched):

  • Fixed in Hotfix 1.000.011 (Sept 24): the developer's patch notes list "Fixed an issue with Genessa's Faithful Doubles when using Lost Clotstone". Update to the latest patch.

Note: The Oct 14 hotfix fixes a related but distinct form: "Lost Clotstone for Genessa Clones using Obsidian Hammer getting stuck and staying forever" — see new entry #51. This entry remains the 1.000.011 Faithful-Doubles form.

Sheep attacks continuing after transforming back

Symptom: Sheep attacks continuing after transforming back — the player kept attacking in sheep form after returning to their shell

Cause: Transformation state bug. The developer has not described the underlying cause.

Fix (developer-patched):

  • Fixed in Hotfix 1.000.011 (Sept 24): the developer's patch notes list "Fixed the player continuing sheep attacks after transforming back into their shell". Update to the latest patch.

Enemies aggroeing during miniboss fights

Symptom: Enemies aggroeing during miniboss fights — enemies could pull aggro during miniboss encounters when they should not

Cause: AI aggro bug. The developer has not described the underlying cause.

Fix (developer-patched):

  • Fixed in Hotfix 1.000.011 (Sept 24): the developer's patch notes list "Fixed enemies becoming aggroed during miniboss fights". Update to the latest patch.

Newly tracked this round (October, pending)

Shell-swap state anomalies

Symptom: Shell-swap state anomalies — invulnerability after switching shells at a beacon (persists until resting again); passive effects possibly double-applying after a shell swap

Report: Community-reported: two players on different platforms (PC Steam, Oct 6; PS5, Oct 4) report state anomalies after swapping shells at a beacon — invulnerability with Proxima after switching from Lazlo, and a roughly doubled Moonshine passive after Lazlo→Gragu→Lazlo. The PS5 reporter speculates passive effects in general may be stacking or re-applying on shell change; this is a player hypothesis, not a confirmed mechanic. No developer response in either thread as of Oct 13.

No workaround yet. This entry is pending a developer response — no player-side fix has been reported.

Shell-unlock achievements not popping

Symptom: Shell-unlock achievements not popping — 'Down with the Thickness' (Lazlo), 'My Brether' (Smert), and 'The Alchemist' stay locked while 'Shell Seeker' (unlock all shells) is unlocked

Report: Community-reported (Oct 2–11, all after the Sept 24 hotfix, version 1.000.011): three players report shell-unlock achievements never triggering, in each case while the unlock-all-shells achievement DID trigger — a self-inconsistent state. The developer replied in two of the threads asking for details (Slayer Seal usage, admin/compatibility mode, Steam overlay, Linux/experimental build) but has not confirmed a bug and is still gathering info. One player ruled out the Slayer Seal and the Proton compatibility layer via a fresh playthrough on the same machine (community-reported self-test).

No workaround yet. This entry is pending a developer response — no player-side fix has been reported.

Trials not spawning enemies

Symptom: Trials not spawning enemies — activated trials spawn no enemies (or no golem while lesser enemies spawn normally), and trial barriers fail to hold the player, making trials impossible to finish

Report: Community-reported (Oct 10): one player reports multiple trial spawn failures across two locations, plus non-functional trial barriers in the same trials; reloading the area or returning to the main menu restored spawns for them (player-tested, not confirmed). No developer response.

Suggested workaround (unverified):

  • Reload the area (or return to the main menu) — player-reported to restore trial spawns; unconfirmed workaround

Smert invincibility via Last Vow + kick spam

Symptom: Smert invincibility via Last Vow + kick spam — spamming kicks during Last Vow can make Smert invulnerable to nearly all damage indefinitely

Report: Community-reported: four independent players report the trigger across Aug 24 to Oct 4 (Steam Discussions threads 582806239606684240, 581680955258998925, 581680955259060645, 581684201210798848) — one reports the trigger as Last Vow plus two kicks, another confirms kick spam alone sustains it, and one reports the invulnerability persisting into October, after multiple interim patches. A developer team member responded to the earliest thread on Aug 25: "I have logged this for the team" (logged only — no confirmation and no fix named as of the Oct 14 hotfix notes).

Suggested workaround (unverified):

  • Return to the main menu or die to an environmental hazard — player-reported to clear the invincibility (unconfirmed)

Night-mode Collector trade UI gone + input lockup

Update (Oct 16): a PS5 patch fixes the Collector softlock; PC and Xbox patches are coming shortly, per the developer. Status unchanged — this page follows the PC patch.

PS5 players: the Oct 16 patch notes list "Fixed softlock when interacting with the collector" — update your game and re-test. PC and Xbox players: no patch on your platform yet; the developer says it is coming shortly. Day-mode trading remains the reliable path in the meantime.

Symptom: Night-mode Collector trade UI gone + input lockup — during night mode, talking to The Collector and selecting Trade makes the merchant UI fail to appear; the dialogue/HUD disappears and controls stop responding, forcing an Alt+F4. Day mode is unaffected.

Report: Two layers. Developer: acknowledged in two threads on Oct 15 — "We are aware of this issue with some merchants since the hotfix and the team are investigating this issue!" and "Yes, can confirm we are currently looking into this issue." — a hotfix-regression merchant issue under active investigation; no root cause or fix named yet. Community-reported: 12 independent players across 5 threads (English, Russian, and Chinese; Oct 13–15, all after Hotfix 1.000.012) describe the identical symptom form: trade menu never opens, UI disappears, input dies, Alt+F4 required. Status update (Oct 16, developer announcement on Reddit, level 4): the developer has released a PS5 patch that names this fix directly — "Fixed softlock when interacting with the collector" — so the softlock half of this issue is patched on PS5. The same announcement says Xbox and PC patches are coming shortly. Status stays investigating: PC is this page's primary audience, the PC patch has not shipped at time of writing, and the fix line names the softlock form only — whether the night-mode trade-UI form (UI never appears) is covered by the same patch has not been stated. Per-platform fix claims are not adopted into the status field until the PC patch ships or a second source confirms coverage.

No workaround is confirmed yet; day-mode trading still works, and the developer is investigating.

Note: fix=null by design: no confirmed workaround exists. Players tried un-equipping Egon's Stone, switching shells/stones, day-night cycling via Festus, restarts, file verification, and resolution/overlay changes — all community-reported, all ineffective. Page copy carries the workaround-watch sentence: "No workaround is confirmed yet; day-mode trading still works, and the developer is investigating." Provenance threads: 564796605131212715 / 564796605131169935 / 564796605131130393 / 564796605131220092 / 564796605131305693. Developer replies verbatim in r12-evidence-round7/th_564796605131212715.txt (#3) and th_564796605131169935.txt (#10). Oct 16 PS5 patch (developer announcement on Reddit; Steam channels carried no announcement at verification time — the patch has not gone through the Steam news feed): the notes list "Fixed softlock when interacting with the collector" verbatim. Two-layer status: PS5 — fix published (per-platform, community confirmation of the fix landing still pending); PC and Xbox — pending, the developer says patches are coming "shortly". This entry is not flipped to fixed: the PC patch is not out, the announcement names no version number (version stays as-is), and the symptom split between the softlock form and the missing-UI form is unresolved. Day-mode trading still works as the interim path on all platforms. Watch item: when the PC patch lands, re-verify against its notes before any status change — see changelog entry ps5-hotfix-1016.

Fixed in Hotfix Update 1.000.012 (Oct 14)

Slayer Seal / Punch issues

Symptom: Slayer Seal / Punch issues — several issues affecting the Slayer Seal and Punch, as described by the developer

Cause: Bugs in the Slayer Seal and Punch. The developer used generalized wording (\"several issues\") and did not enumerate the affected forms or describe the underlying cause.

Fix (developer-patched):

  • Fixed in Hotfix Update 1.000.012 (Oct 14): the developer's patch notes list \"Fixed several Slayer Seal/Punch issues\". Update to Hotfix Update 1.000.012 (Oct 14).

See also: Distinct from #18 (Slayer Seal UI lockup, fixed HF2.0; re-fixed 1.000.011 unresponsive-seal form) — different symptom forms, #18 stays unchanged.

Mether's Breath second-warning soft lock

Symptom: Mether's Breath second-warning soft lock — backing out of Mether's Breath's second warning soft-locked the game

Cause: UI prompt/interaction bug. The developer has not described the underlying cause.

Fix (developer-patched):

  • Fixed in Hotfix Update 1.000.012 (Oct 14): the developer's patch notes list \"Fixed backing out of Mether's Breath's second warning soft-locking the game\". The same notes add \"Mether's Breath now has multiple warning prompts\". Update to Hotfix Update 1.000.012 (Oct 14).

See also: Distinct from #33 (Mether's Severance soft lock, fixed 1.000.011) — different ability, different form. Same soft-lock entry pattern as #33 precedent.

Genessa Clones stuck with Lost Clotstone (Obsidian Hammer)

Symptom: Genessa Clones stuck with Lost Clotstone (Obsidian Hammer) — Genessa Clones using Obsidian Hammer with Lost Clotstone got stuck and stayed forever

Cause: Item-interaction bug. The developer has not described the underlying cause.

Fix (developer-patched):

  • Fixed in Hotfix Update 1.000.012 (Oct 14): the developer's patch notes list \"Fixed Lost Clotstone for Genessa Clones using Obsidian Hammer getting stuck and staying forever\". Update to Hotfix Update 1.000.012 (Oct 14).

See also: Cross-reference #39 (Genessa's Faithful Doubles + Lost Clotstone, fixed 1.000.011): same item family, distinct symptom form — that fix covered the ability behaving incorrectly; this covers stuck clones persisting forever. Kept as a separate entry per recommendation; #39 gets a see-also note, not a merge.

Gragu possession blocked

Symptom: Gragu possession blocked — Gragu could not be possessed

Cause: Possession logic bug. The developer has not described the underlying cause.

Fix (developer-patched):

  • Fixed in Hotfix Update 1.000.012 (Oct 14): the developer's patch notes list \"Fixed issue that blocked Gragu from being possessed\". Update to Hotfix Update 1.000.012 (Oct 14).

See also: Distinct from #22 (Gragu barkeep softlock, fixed Week 1) — different form. #22 gets a distinction note; not merged.

These reports carry their stated evidence level. Unconfirmed community claims are tracked, and no unconfirmed numbers are stated as fact. Entries move to the triage phases above once confirmed by a second source or a developer response — and entries the developer has patched carry Fixed status with the patch note quoted.

Common crashing patterns across all phases

After documenting the triaged issues, a few patterns stand out that are worth understanding even before you dive into specific fixes. The single most important pattern: D3D12 GPU crashing is the most common crashing type in Mortal Shell 2, and it's the root cause behind roughly a third of all crashing reports. These show up as "GPU Crash dump Triggered" or "D3D device being lost" errors, and they're exacerbated by GPU overclocking, outdated drivers, and Frame Generation. Hotfix 1.0 (August 17) targeted these crashing issues directly, and Hotfix 2.0 (August 18–19) addressed further instances, but some players still report D3D12 crashing even on clean driver installs — suggesting the engine itself still has edge cases.

The second pattern is Frame Generation as a universal crashing amplifier. DLSS Frame Generation is the number one trigger for crashing in Mortal Shell 2. It doesn't just cause one specific crashing issue — it amplifies at least three separate crashing types (death/reload crashing, mid-game D3D12 crashing, and GC exception crashing). If you're crashing and Frame Generation is on, turn it off and restart. That single step resolves crashing for a significant portion of affected players.

The third pattern involves shader compilation crashing at launch. Multiple crashing issues cluster around the game's first-run shader compilation step. The Oodle decompression error during shader compilation, the black screen after shader compile, and post-death shader cache exhaustion all share a common thread: the UE5 shader pipeline is fragile when cache files are corrupted, undersized, or stored on slow media. Setting your NVIDIA Shader Cache Size to Unlimited and installing the game on an SSD addresses most of these.

Finally, garbage collection crashing appears in at least three separate issues (the GC exception crashing, the settings menu GC crashing, and the Burn status crashing, which also traces through GarbageCollection.cpp). UE5's GC system is more aggressive than UE4's, and certain in-game interactions — opening the settings menu, applying the Burn status effect with specific weapon combinations, or running the game on certain Intel 13th/14th-gen CPUs with known instability issues — can push it over the edge, triggering crashing. Two of these have workarounds; one (the GC exception crashing) remains pending.

One more pattern worth understanding: the render-path GPU crash cluster. Three lines of evidence currently point at the same rendering pipeline. First, VFX-heavy scene degradation and crashes after the September 5 Adaptive Difficulty patch (see the community-reported entry below). Second, the random D3D12 GPU crash dump (Phase 2, above) — players report it did not occur during advanced access and started appearing after the early hotfixes, with new reports continuing on September 9–10 after the Adaptive Difficulty patch; the developer confirmed on Steam Discussions (August 25 and August 28) that it is "actively working on this issue and are looking into a fix." Third, a hardware-side elimination thread on Reddit (different hardware configurations all reproduce it, which argues against a single-hardware cause) — we have not independently verified that thread, so treat it as an unverified lead, not a fact. These three lines are independent but consistent: if your crashing matches this cluster, the D3D12 workarounds in Phase 2 are the best current mitigation while we wait on the developer's fix.

If you've spent any time searching for "Mortal Shell 2 crashing fixes," you've probably noticed that the first several pages of results are dominated by SEO farm sites — the kind that auto-generate guides by scraping Reddit threads, paraphrasing Steam forum posts, and padding them with generic advice like "update your drivers" and "run as administrator." These sites don't test anything themselves. They don't reproduce error strings. They don't distinguish between a D3D12 GPU crash and a garbage collection exception — to them, all crashing is the same. Worse, some of the advice they circulate is actively risky: disabling Windows security features, modifying system-level GPU settings without understanding the tradeoffs, or applying launch options that haven't been verified against the actual error. Below is advice we've seen circulating on these sites that we could not verify through community testing — or that carries real risk if applied incorrectly. We've noted why each one is suspect.

Crash on load: fixes for every loading-screen scenario

"Crash on load" covers three different failure points, and each has its own entry:

  • Crash when pressing Continue at the main menu — the game consistently crashes on first load before reaching gameplay. Community-reported by three independent players since late August; the developer confirmed at the crash-cluster level that they are "aware of these crashing issues and are looking into a solution," but has not named this specific symptom. Whether Hotfix 1.000.011 addressed it is unconfirmed. Reporter-verified workarounds: a 30 FPS lock before pressing Continue, or loading a fresh save slot first (both must be re-applied after each restart). [fixes #45]
  • UE error when loading a save — fixed in the Week 1 Update. Still crashing after updating? Verify game files, then report it to the developer. [fixes #12]
  • Freeze or crash during "Compiling Shaders" before the main menu — an Oodle decompression error; setting the NVIDIA Shader Cache Size to Unlimited is recommended by the Playstack community manager. [fixes #1]

Mortal Shell 2 freezing: the four known lockup patterns

  • Completely locks up after death — loading screen never resolves. Cause tracks to shader cache exhaustion or D3D device loss; the reported fix is NVIDIA Shader Cache Size → Unlimited. [fixes #13]
  • Freezes during "Compiling Shaders" — see the shader compilation entry above; retrying compilation multiple times is part of the reported workaround. [fixes #1]
  • Black screen with audio still playing after shader compile finishes — a developer-acknowledged known issue with no confirmed fix yet; a clean driver reinstall, file verification, and disabling overlays are the suggested things to try. [fixes #10]
  • Game closes after long play sessions — a UE5 memory leak; the reported workaround is a static page file (initial 16384 MB / max 32768 MB on SSD). [fixes #16]

What error code 0x11C7 means (and why the game is safe to run)

If Windows 11 blocks the launch with "An Application Control policy has blocked this file (0x11C7)", that is Smart App Control false-positive flagging MortalShell2.exe — the developer replied on Steam Discussions (Aug 25) that the game is safe to run and the flag is unnecessary. The developer-recommended step is turning off Smart App Control or making an exception for the game; note that Windows policy only allows Smart App Control to be fully disabled once per install. The developer has logged the issue. [fixes #31]

"GPU Crash dump Triggered" (FD3D12DynamicRHI::TerminateOnGPUCrash)

This DX12 driver crash was targeted by Hotfix 1.000.010, but some players still report it. The developer confirmed on Steam Discussions (Aug 25 & Aug 28) that they are "actively working on this issue and are looking into a fix." New reports of the same crash dump continued after the September 5 Adaptive Difficulty patch. A community analysis traced one dump to INVALID_POINTER_READ in SetShaderParametersOnContext — unconfirmed. [fixes #5]

EXCEPTION_ACCESS_VIOLATION codes — which one do you have?

  • Reading 0xffffffffffffffff after dying, tied to Frame Generation — disable FG, switch to Fullscreen, drop to High preset, FSR → TSR. [fixes #3]
  • 0x00000000000002c8 in a GarbageCollection stack — a UE5 GC bug that may correlate with unstable Intel 13th/14th-gen CPUs; disable FG + Ray Tracing and cap FPS. Still pending. [fixes #6]
  • 0x0000000000000000 during a burn-status death — the Lazlo Retribution + Burn interaction; avoid the Burn status effect until patched. [fixes #7]

Rule Out Hardware-Side Instability (Intel 13th/14th-Gen)

Some crash signatures in this guide — particularly Oodle shader-decompression errors (issue #1) and GC exceptions on Intel 13th/14th-gen systems (issue #6) — overlap with a known hardware-side topic: instability on some Intel 13th/14th-gen desktop CPUs. This section does not say your crash is caused by your CPU. It is a rule-out checklist: if the steps below change or stop your crashes, the cause is likely outside the game; if crashes continue identically afterward, treat them as game-side and keep reporting them.

  1. Update your BIOS and microcode. Motherboard vendors have shipped BIOS updates carrying Intel's latest microcode patches and Intel Default Settings profiles for 13th/14th-gen chips. Install the newest BIOS for your board and load the Intel Default Settings profile before testing the game again.
  2. Test your memory. Run MemTest86 (or the Windows Memory Diagnostic) for a full pass. Memory errors can surface as exactly the kind of random decompression and garbage-collection crashes described elsewhere on this page.
  3. Reset to defaults. Set your BIOS to optimized defaults (no manual overclock, no undervolt, XMP/EXPO can be toggled off for one test run) and re-test. An unstable memory or core configuration can mimic game bugs.
  4. If crashes persist through steps 1–3, contact Intel or your system/board vendor about warranty service (RMA). Intel's 13th/14th-gen warranty coverage for affected boxed processors has been extended by two years per Intel's own communications — confirm the current terms with Intel directly before filing.

For reference: one Steam player with an i7-14700K reported this game's Oodle shader crash signature persisting after the Week 1 patch (community-reported, Aug 30; their post-Sept-24 status is unknown), and the Oodle error text in that report links to Oodle's own advisory pointing toward Intel 13th/14th-gen instability — a vendor-side attribution, not a confirmation by the game's developer. A separate Linux/NVIDIA player's Xid 13 logs (community-reported, measured by one player) showed the fault rate dropping sharply at lower frame caps; see issue #6.

Crashing FAQ

These are the most common questions players ask about crashing in Mortal Shell 2, distilled from Steam forums, the game's Discord server, and the 52 issues tracked on this page. If your specific question isn't here, check the crash entries above — the symptom → cause → fix format is designed to be searchable.

Why does Mortal Shell 2 crash on startup?

Most startup crashes trace to shader compilation — set NVIDIA Shader Cache Size to Unlimited, clear DXCache/GLCache, or retry compilation. See Phase 1.

Does turning off Frame Generation stop crashes?

It resolves the most common mid-game and respawn crashes — Frame Generation conflicts with the UE5 render pipeline, especially 2x/3x auto modes.

Are crashes fixed after Hotfix 2.0?

Partially. Hotfix 2.0 fixed the Slayer Seal UI lock and the Riposte weapon steal, and the Week 1 Update fixed the save-load crash plus three progress softlocks — Beacon fast-travel crashes remain pending.

Is -UseFixedTimeStep a safe fix?

No — it is unverified advice circulating on content-farm sites. We list it under Not Recommended.

Where are Mortal Shell 2 saves located?

Under %AppData%\Local\MortalShell2\Saved\SaveGames — deleting the EnhancedInputUserSettings and SpartaGameSettings files fixes the keybind menu crash.

Why does Mortal Shell 2 keep crashing during gameplay?

In-game crashes cluster around three tracked causes: Frame Generation conflicting with the UE5 render pipeline (the most common trigger — see Phase 1), D3D12 GPU crash dumps from driver or overclock instability, and EXCEPTION_ACCESS_VIOLATION errors from the UE5 garbage collector. VFX-heavy scenes have also degraded performance since the September 5 Adaptive Difficulty patch. Work through the phases in order and stop when the crashing stops.

Why does Mortal Shell 2 crash on load?

Crashing on load splits into three tracked scenarios: a crash when pressing Continue at the main menu (still pending — community-reported workarounds are a 30 FPS lock before Continue or loading a fresh save slot first), shader-compilation crashes before the main menu (set NVIDIA Shader Cache Size to Unlimited), and the UE save-load error that the Week 1 Update already fixed — if you still see it, verify game files first. Details in the Crash on load section above.

What if none of these fix my crashing?

If you've worked through all three phases and you're still crashing in Mortal Shell 2, you're in a smaller but real group of players. Here's what to do next, in order of usefulness:

  1. Check the three pending crashes. If your crashing matches the Beacon teleport crash (crashing when interacting with or leaving a Beacon), the black screen after shader compilation (audio plays but no picture), or the GC exception crash (EXCEPTION_ACCESS_VIOLATION in a GarbageCollection stack), there is no confirmed player-side fix right now. These three crashes are acknowledged issues that require a developer patch. The developer has shipped eight patches since launch (Hotfix 1.0 on August 17, Hotfix 2.0 on August 18–19, Balance Patch 1 on August 22–23, the Week 1 Update on August 29, Hotfix 3 on September 1, the Adaptive Difficulty update on September 5, Hotfix 1.000.010 on September 14, and Hotfix 1.000.011 on September 24), so more fixes are likely coming.
  2. Collect your crash log. Mortal Shell 2 writes crash dumps to %LocalAppData%\MortalShell2\Saved\Crashes. The folder name and the callstack inside it are the fastest way to identify which of the 52 issues you're hitting. If the error string matches something on this page but the listed fix isn't working for you, your specific hardware or driver version may need a different approach. Newly reported issues without a confirmed fix are collected under "Recently reported (community)" below the phases; entries the developer has since patched carry Fixed status in that section.
  3. Report it with details. The developer's bug reporting channels (Steam discussions and the game's Discord) need specific information to act on: your GPU/CPU/RAM, driver version, the exact crash log, and what you were doing in-game when the crashing started. "Game keeps crashing" without error strings or reproduction steps won't get traction. The three pending crashes got developer attention specifically because players reported them with error strings and consistent reproduction steps.
  4. Consider a refund window. If crashing makes the game unplayable and none of the 22 documented issues apply to your situation, remember that Steam's refund policy covers games with under two hours of playtime. If you're past that window, you can still contact Steam Support with your crash logs — they've been known to make exceptions for games with widespread technical issues.
  5. Check back after the next patch. The developer has been releasing patches at a rapid pace. Twenty-one of the 52 tracked issues have been fixed by the developer so far — the Week 1 Update resolved the save-load crash and three progress softlocks, Hotfix 1.000.011 (September 24) developer-patched ten recently reported items in one pass, and Hotfix 1.000.012 (October 14) added four more (the Slayer Seal/Punch issues, the Mether’s Breath second-warning soft lock, the Genessa Clones Lost Clotstone stuck form, and blocked Gragu possession). If your crashing is caused by one of the three pending issues, the next patch cycle is your best bet. This page is updated within 24 hours of every patch release.

Crashing right at the Continue button before gameplay loads? That matches a separately tracked first-load crash reported by multiple players — try the 30 FPS lock or the new-save-slot workaround in the quick-fix area above, and note both need re-applying after every restart. It is listed as pending: the developer has acknowledged the surrounding crash cluster but has not confirmed a fix for this specific crash yet.

And once the crashing stops for good, the fun part starts: pair your shell with the weapons it was meant for in the Build Planner — four steps, synergy notes included, no account needed.

Related