---
id: "GHSA-wf6j-7hvr-gr93"
kind: "vulnerability"
aliases: ["GHSA-wf6j-7hvr-gr93"]
title: "Hydra legacy target blacklist permits unsafe PyYAML loading"
severity: "high"
cvss: 7.8
epss: null
exploited: false
status: "open"
category: "package"
products: [{"name": "hydra-core", "fixed": ["1.3.8"], "affected": ">= 1.3.6, < 1.3.8", "ecosystem": "pypi"}, {"name": "hydra-core", "fixed": ["1.4.0rc3"], "affected": ">= 1.4.0.dev9, < 1.4.0rc3", "ecosystem": "pypi"}]
ecosystems: ["pypi"]
tags: ["cwe-184", "hydra-core", "ml-pipelines", "package", "python", "pyyaml"]
not_affected_if: ["default-config", "trusted-config-only"]
published_at: "2026-10-27T17:08:41+00:00"
first_seen_at: "2026-10-27T17:11:59+00:00"
updated_at: "2026-10-27T17:13:15+00:00"
sources: [{"source": "ghsa", "url": "https://github.com/advisories/GHSA-wf6j-7hvr-gr93", "first_seen_at": "2026-10-27T17:11:59+00:00"}]
---

# Hydra legacy target blacklist permits unsafe PyYAML loading

Hydra's legacy instantiate() blacklist omits PyYAML's unsafe loading functions, so a config-controlled target such as yaml.unsafe_load can run Python code when an application passes untrusted targets and arguments into it. The configured callable runs with the application's process privileges, and the same pattern can arise from a third-party model whose bundled configuration is loaded by the app.

_Summary, check and fix written by a model from the sources; labelled as such._

## Affected

- pypi:hydra-core affected >= 1.3.6, < 1.3.8 fixed 1.3.8
- pypi:hydra-core affected >= 1.4.0.dev9, < 1.4.0rc3 fixed 1.4.0rc3

## How to check

- Run `pip show hydra-core` and check whether the version is `1.3.6` or `1.3.7`, or a 1.4 prerelease `1.4.0.dev9`, `1.4.0.dev10`, `1.4.0rc1` or `1.4.0rc2`.
- Search application code for calls to `instantiate(` or `call(` from hydra and check whether any config passed to them comes from users, uploads or downloaded model metadata.

## Fix

- Upgrade hydra-core to `1.3.8` on the 1.3 line.
- Upgrade hydra-core to `1.4.0rc3` on the 1.4 line.
- Until upgraded, do not pass untrusted configuration to `instantiate()` or `call()`; for Hydra 1.4, supply a narrow execution whitelist from trusted Python code.

## Timeline

- 2026-10-27T17:11:59+00:00 ghsa https://github.com/advisories/GHSA-wf6j-7hvr-gr93

## From the sources

## Summary

Hydra's legacy target blacklist omits PyYAML's unsafe loading surfaces. If an
application passes attacker-controlled targets and arguments to instantiate(),
a target such as yaml.unsafe_load can execute Python code through a
config-controlled YAML stream containing Python object-construction tags.

## Impact and prerequisites

The attacker must control the relevant target and arguments that the application
instantiates. The configured callable executes with the application's process
privileges. This is a gap in the best-effort legacy blacklist. It is not a bypass
of Hydra 1.4's execution whitelist supplied by trusted Python code.

One realistic exposure is a model downloaded from a third-party repository,
such as Hugging Face, that bundles attacker-controlled configuration. When a
victim loads the model using application code that passes this configuration
to legacy Hydra instantiation without effective target restrictions, the
model's metadata can execute code in the victim process. This model-loading
attack pattern is documented in the original Hydra advisory
GHSA-2cp2-2r3c-7p7r and the associated Unit 42 research. It does not imply that
all Hugging Face models or currently patched downstream loaders are affected.

Hydra 1.3's blacklist is defense in depth, not a complete security boundary.
Hydra 1.4 is affected on its deprecated no-whitelist legacy path. Installed
application wrappers can perform indirect operations beyond a name-based
blacklist's coverage; this fix does not make arbitrary configuration safe.

The unsafe_load finding is directly reproduced. PyYAML 6.0.3 full_load rejects
the demonstrated python/object/apply payload, so the finding does not claim that
all PyYAML loading APIs execute that payload.

## Remediation

The fix blocks the yaml. target namespace except for the exact yaml.safe_load,
yaml.safe_load_all, yaml.safe_dump, and yaml.safe_dump_all functions. Loader
classes, constructor registration, general dump/load APIs, and other targets in
the namespace remain blocked. Existing canonical-identity checks also cover
aliases. The fix preserves Hydra 1.4's explicit unsafe opt-out for trusted use.

## Workarounds

Do not pass untrusted configuration to instantiate() or call(). For Hydra 1.4,
supply a narrow execution whitelist from trusted Python code. Do not let
untrusted configuration select its own execution policy or opt out of checks.

## Affected releases

The reported incomplete-remediation stable range is hydra-core 1.3.6 through
1.3.7. The 1.4 prerelease legacy path is also affected, including 1.4.0.dev9,
1.4.0.dev10, 1.4.0rc1, and 1.4.0rc2. The related blacklist advisory
GHSA-rqx7-p7vv-w7hr was first patched in 1.3.6 and 1.4.0.dev9; these ranges
describe the remaining omission after that remediation.

## Credit

Reported by zx (@manus-pi).

---

_Free to read at a human pace. An API with keys and an MCP server are in preparation; early access: hello@infrastack.ai_