Skip to content

Security: ecrl/padelpy

Security

Report a vulnerability

SECURITY.md

Security policy

Supported versions

Security fixes are considered for the latest published release on PyPI and for the active modernization line targeting 0.1.17. Older patch lines may not receive backports.

Reporting a vulnerability

Please do not open a public GitHub issue for security vulnerabilities.

Prefer GitHub’s private vulnerability reporting for ecrl/padelpy. That path creates a private advisory draft maintainers can triage without public disclosure.

Alternatively, report by email to travis.j.kessler@gmail.com with:

  • A description of the issue and its impact
  • Steps to reproduce, or a proof of concept if available
  • Affected versions / commit hashes if known

You should receive an acknowledgment within a few business days. We will coordinate a fix and disclosure timeline as appropriate.

Threat model

padelpy is a thin local wrapper. Typical use is a researcher or pipeline invoking descriptor calculation on inputs they already control. It is not a network service and does not expose an HTTP API.

Trust boundaries

Boundary Assumptions
Caller Controls SMILES/MDL/SDF content and optional filesystem paths passed to padeldescriptor (config, mol_dir, d_file, descriptortypes, tautomerlist). Path arguments are forwarded to the PaDEL CLI as local trusted inputs, not as a sandbox.
Wrapper Builds a Java argv list (no shell), writes temporary structure files under a process-private temporary directory, and optionally enforces a subprocess timeout.
System Java The java executable on PATH is trusted. A compromised or unexpected JRE is outside padelpy’s control.
Vendored PaDEL/CDK Bundled JARs under src/padelpy/PaDEL-Descriptor/ execute as part of descriptor calculation. Bytecode behavior and upstream library CVEs may affect the process; see below.

Likely vs unlikely risks

  • In scope for padelpy: wrapper command construction, temporary-file handling, packaging/supply-chain of the Python distribution, and publish CI integrity.
  • Often upstream / environmental: PaDEL hangs or crashes on particular structures; CVEs in Guava/CDK/other bundled libraries that require attack patterns PaDEL does not expose (for example remote deserialization); a malicious or mismatched system java.
  • Caller responsibility: Do not pass untrusted path arguments into padeldescriptor in multi-tenant or web-facing services without an outer sandbox. Prefer from_smiles / from_mdl / from_sdf with timeouts when wrapping user-supplied chemical structures.

Issues that are solely upstream PaDEL/CDK behavior may be documented as limitations rather than treated as padelpy vulnerabilities; still report them if you believe they create a security risk through this wrapper.

Vendored Java artifacts

Each wheel and sdist ships the PaDEL-Descriptor tree (approximately 23 MB), including third-party JARs such as CDK and Guava. Descriptor numeric parity for the 0.1.x compatibility series depends on this frozen bundle; upgrades require an approved dual-run parity plan (see docs/design/DESIGN.md).

Integrity

SHA-256 digests for bundled jars and supporting data files are recorded in docs/security/vendored-artifacts.sha256. Paths are relative to the repository root. Verify a checkout with:

shasum -a 256 -c docs/security/vendored-artifacts.sha256

Regenerate after any intentional change to the vendored tree:

find src/padelpy/PaDEL-Descriptor -type f \( -name '*.jar' -o -name '*.xml' -o -name '*.xls' \) \
  | sort \
  | xargs shasum -a 256 \
  > docs/security/vendored-artifacts.sha256

Advisory posture

Some bundled libraries are old relative to current upstream releases (for example Guava 17.0). Public advisories may list those versions as affected even when a practical exploit path through the PaDEL CLI is unlikely. Until a parity-gated JAR modernization lands, treat such findings as known supply-chain debt: report them privately if you have evidence of impact via padelpy, and prefer process isolation when running padelpy on untrusted chemical inputs in shared environments.

Python dependency auditing

Runtime padelpy installs have no Python dependencies beyond the standard library. Development and CI use pip-audit on the installed environment (see the audit job in .github/workflows/ci.yml). Run locally after pip install -e ".[dev]":

pip-audit --strict

There aren't any published security advisories