Versioned security assessment

Report ID: SA-5E044157

6/30/2026, 5:49:39 AM

diffdock security assessment v5

Skill Security Certification Report

Audit History
Audit model: codex Historical report
Skill name
diffdock
Version
v5
Maintainer
K-Dense-AI
Source
github.com/K-Dense-AI/claude-scientific-skills/tree/main/scientific-skills/diffdock
Coverage
9 Files scanned · 2,188 Lines analyzed
Policy version
Unavailable

Highest confirmed finding severity

Medium

2 confirmed security findings require 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 many critical and high alerts, but review found they are mostly false positives from Markdown command examples, molecular terminology, amino acid sequences, and local scientific helper scripts. The remaining real risk is operational: users are guided to clone external repositories, pull Docker images, run shell commands, launch a local web UI, and write local output files. No evidence found of credential theft, prompt injection, obfuscation, command injection, or malicious data exfiltration.

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 · 2,188 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.

  1. Source

    Binding unavailable

  2. Artifact

    Identity incomplete

  3. Audit

    Complete

  4. 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.

Observed in 4 evidence locations

Network access

May connect to external services.

Observed in 5 evidence locations

Filesystem access

May read or write local files.

Observed in 4 evidence locations

Env variables

May read values from the process environment.

Not recorded by this audit

External commands

May invoke commands or programs outside the Skill.

Observed in 5 evidence locations

Risk findings

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

Confirmed security concerns (2)

RISK-001 Medium
SKILL.md:52-63references/workflows_examples.md:9-26SKILL.md:363-371
External Commands and Third-Party Downloads Required
TRUE_POSITIVE with legitimate intent. The skill instructs users to clone DiffDock, create Conda environments, pull and run a Docker image, run Python modules, and optionally use an online demo. These commands are normal for DiffDock but require users to trust external repositories, images, and downloaded model assets.
The commands and URLs are explicit and semantically tied to installing and running DiffDock. The risk is operational trust and execution risk, not evidence of malicious behavior.
RISK-002 Medium
scripts/prepare_batch_csv.py:76-82scripts/prepare_batch_csv.py:177-178scripts/analyze_results.py:95-109scripts/analyze_results.py:214-216
Local File Reads and Writes in Helper Scripts
TRUE_POSITIVE with limited scope. The included scripts read user-supplied CSV and result directories, validate referenced file paths, and write CSV summaries or templates. This is expected scientific workflow behavior, but users should run it only on intended local paths and review output targets.
The filesystem operations are directly visible and have normal CLI-controlled scope. I did not find deletion, credential harvesting, or network transmission of file contents.

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 (3)
Low
SKILL.md:81-88references/parameters_reference.md:157-160references/workflows_examples.md:39-45
Static Command Execution Alerts Are Mostly Documentation Examples
FALSE_POSITIVE for command injection. Most external command alerts point to fenced Markdown examples or parameter names that show users how to run DiffDock, GNINA, Docker, and helper scripts. They are not hidden execution paths inside the skill runtime.
The flagged lines are Markdown documentation or explicit user-run commands. I did not find code that constructs shell commands from untrusted input.
Low
assets/batch_template.csv:3scripts/prepare_batch_csv.py:172references/confidence_and_limitations.md:3
Sensitive Data and Weak Cryptography Alerts Are Domain False Positives
FALSE_POSITIVE for credential and cryptography risk. The SAM hit is an amino acid sequence containing the substring SAM, and the weak-crypto hits map to molecular docking text, filenames, or model terminology rather than cryptographic code.
The reviewed lines contain biological sequences or natural-language documentation. No evidence found of Windows SAM access, hashing implementation, or weak cryptographic use.
Low
scripts/setup_check.py:34-46scripts/setup_check.py:92-109
Dynamic Import Is Restricted to Dependency Checks
FALSE_POSITIVE for malicious dynamic loading. The setup checker calls __import__ only with package names from hardcoded dependency lists to report installation status. It does not import user-provided module names.
The import target comes from fixed dependency tuples in the script. No user-controlled import path or plugin loading behavior was found.

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