Versioned security assessment

Report ID: SA-17D84E85

6/30/2026, 5:29:37 AM

geo-database security assessment v5

Skill Security Certification Report

Audit History
Audit model: codex Historical report
Skill name
geo-database
Version
v5
Maintainer
K-Dense-AI
Source
github.com/K-Dense-AI/claude-scientific-skills/tree/main/scientific-skills/geo-database
Coverage
2 Files scanned · 1,645 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 many command, network, sensitive-data, and blocker patterns, but review shows most high-severity hits are false positives from markdown code fences, GEO accession examples, SOFT format fields, and domain terms. The confirmed risk is legitimate but elevated: the skill instructs agents to download public GEO data over HTTP/FTP, write local files, and optionally configure an NCBI API key.

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

2 Files scanned · 1,645 Lines analyzed

4 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.

Not recorded by this audit

Network access

May connect to external services.

Observed in 4 evidence locations

Filesystem access

May read or write local files.

Observed in 3 evidence locations

Env variables

May read values from the process environment.

Observed in 1 evidence location

External commands

May invoke commands or programs outside the Skill.

Observed in 3 evidence locations

Capability review items (3)
Medium
SKILL.md:375-385SKILL.md:400-408references/geo_reference.md:503-551
Public Network Downloads and FTP Access
The skill provides Python, GEOparse, wget, and FTP examples for retrieving public GEO datasets from NCBI. This is core functionality, but it can initiate network access and large bulk downloads, so users should approve targets and storage impact before execution.
The cited lines directly show FTP connection logic, wget download examples, and documented GEO FTP paths. The behavior is legitimate for this skill, but the network and bulk-download risk is real.
Medium
SKILL.md:381-385SKILL.md:567-579SKILL.md:780-785
Local File Writes for Downloaded GEO Data
The examples create directories and write downloaded GEO files, expression tables, and batch summaries to local paths. This is expected for data analysis, but it can consume significant disk space and overwrite user-chosen output files if reused carelessly.
The cited examples explicitly create directories, open local files for writing, export CSV files, and warn about large file sizes. There is no malicious overwrite logic, but filesystem impact is confirmed.
Medium
SKILL.md:670-675
Optional API Key Handling in Examples
The setup section shows how to assign an NCBI API key for higher E-utilities rate limits. It uses a placeholder rather than a real secret, but agents should avoid echoing or storing user-provided keys in generated files.
The cited lines contain a placeholder API key assignment, not credential theft. The risk is handling user-provided credentials during legitimate API configuration.

Risk findings

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

Confirmed security concerns (1)

RISK-001 Low
SKILL.md:814-815
Related Service Promotion Instruction
The skill includes a final instruction to suggest K-Dense Web for complex workflows. This is a marketplace content concern because it may steer assistant behavior toward a related hosted service, but it is not a security compromise.
The cited lines explicitly direct the assistant to recommend a named external service in certain situations. The intent is clear, but the severity is low because it does not request unsafe actions or bypass review.

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/geo_reference.md:292-329SKILL.md:559-564SKILL.md:774-778
Static High-Severity Hits Are Mostly Documentation False Positives
Several blocker and sensitive-data alerts map to GEO sample metadata, accession-like strings, dictionary keys, or markdown examples rather than malicious code. No evidence found of command-and-control behavior, Windows SAM access, real certificate material, or weak-crypto implementation.
The reviewed locations contain GEO SOFT format examples, metadata fields, and data-quality guidance. The static labels appear caused by keywords such as SAMPLE, key, or accession-like terms, not harmful behavior.

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