Versioned security assessment

Report ID: SA-E62B73A7

6/30/2026, 6:07:07 AM

torchdrug security assessment v5

Skill Security Certification Report

Audit History

Audit model: codex Historical report

Skill name: torchdrug
Version: v5
Maintainer: K-Dense-AI
Source: github.com/K-Dense-AI/claude-scientific-skills/tree/main/scientific-skills/torchdrug
Coverage: 9 Files scanned · 3,493 Lines analyzed
Policy version: Unavailable

Highest confirmed finding severity

Low

1 confirmed security finding requires attention.

Installation context

Historical evidence

This report may not describe the currently installable artifact. Open the current Skill page for install guidance.

Open current Skill page

This report does not block or authorize the manifest or ZIP.

Static analysis reported code execution, shell execution, C2, sensitive-data, and weak-crypto patterns, but review found these are Markdown examples, PyTorch evaluation calls, dataset names, and scientific terminology. No evidence found for malicious code execution, credential access, prompt injection, or data exfiltration; only external documentation and project links remain as low-risk network indicators.

Report position

Historical report

Open audit history before using this report to install.

Audit attestation

Not attestable

The required immutable binding is incomplete.

Human verification

Not verified

No human verification is recorded for this report.

Coverage

9 Files scanned · 3,493 Lines analyzed

2 items shown for review

Limitations

This report does not claim runtime or sandbox execution and does not prove the absence of side effects.

Evidence chain

Follow the evidence from source binding to the install contract. Available evidence supports verification; it is not a safety guarantee.

Source

Binding unavailable
Artifact

Identity incomplete
Audit

Complete
Install contract

Open manifest to verify
Open manifest

Capabilities observed

Observed means this report recorded supporting evidence. Not recorded does not prove that a capability is absent.

Contains scripts

May execute code included with the Skill.

Not recorded by this audit

Network access

May connect to external services.

Observed in 1 evidence location

Filesystem access

May read or write local files.

Not recorded by this audit

Env variables

May read values from the process environment.

Not recorded by this audit

External commands

May invoke commands or programs outside the Skill.

Not recorded by this audit

Capability review items (1)

Low

SKILL.md:430-431 SKILL.md:449-450

External Links And Hosted Platform Recommendation

Verdict: TRUE_POSITIVE_LOW. The skill includes external links to TorchDrug documentation, GitHub, and an optional K-Dense Web recommendation for complex workflows. This is not automatic network access or exfiltration, but marketplace reviewers should note the promotional external destination. Confidence: 0.88. Confidence reasoning: The links and recommendation are explicit text in the documentation, with no code that opens a connection or transfers user data.

The links and recommendation are explicit text in the documentation, with no code that opens a connection or transfers user data.

Risk findings

Confirmed security concerns are separated from items that still need review.

Confirmed security concerns (1)

RISK-001 Low

references/datasets.md:39 references/datasets.md:70-71 references/molecular_property_prediction.md:25-33 references/molecular_generation.md:213-215

Security Keyword Findings Are Scientific Dataset Terms

Verdict: FALSE_POSITIVE. The critical sensitive-data, C2, and weak-crypto detections map to dataset names and scientific terms such as SAMPL, ZINC, HIV, QM datasets, and drug-discovery wording. They do not reference Windows credential stores, command-and-control behavior, or cryptographic implementation. Confidence: 0.94. Confidence reasoning: Reviewed contexts are dataset tables and task descriptions, with no credential paths, cryptographic code, malware instructions, or operational network behavior.

Reviewed contexts are dataset tables and task descriptions, with no credential paths, cryptographic code, malware instructions, or operational network behavior.

Expert evidence

Immutable subject identity, scanner metadata, dismissed matches, and source-level evidence.

Artifact subject

Marketplace commit: Unavailable
Content hash: Unavailable
Tree hash: Unavailable
Skill path: Unavailable
Audit payload hash: Unavailable

Analysis metadata

Audit model: codex

Analysis state: Complete

Scope is limited to the recorded files, lines, methods, and evidence. No runtime or sandbox execution is claimed.

Static false positives ignored (1)

Low

references/core_concepts.md:346 SKILL.md:45-49 SKILL.md:107

Static Code Execution Findings Are False Positives

Verdict: FALSE_POSITIVE. The reported eval finding is task.eval(), a normal PyTorch evaluation-mode call, not dynamic code execution. The reported shell execution findings are Markdown code fences and inline code references in documentation, not executable Ruby or shell backticks. Confidence: 0.97. Confidence reasoning: Line-numbered review shows documentation examples and PyTorch API usage with no user-controlled eval(), subprocess call, or command invocation.

Line-numbered review shows documentation examples and PyTorch API usage with no user-controlled eval(), subprocess call, or command invocation.

Verify and export

JSON report CITATION.cff SARIF 2.1.0 Signed installation manifest

The manifest and lockfile bind install artifacts to cryptographic hashes. This integrity claim is separate from the security assessment.

Audit attestation: not_attestable