Geoskill: Cropland Change Compliance

• Outcome: Review
• Latest audit: Jul 31, 2026 · 12:05 PM UTC
• Version: 2.0.0

Overview

This skill has a real review concern because its remote-download mode can treat ordinary satellite preview images as NDVI data and produce compliance reports from mismatched inputs.

Review this skill before installing in any workflow where compliance reports could influence real decisions. Use validated local NDVI rasters rather than the bbox/date auto-download mode unless the downloader is fixed to compute or verify NDVI from proper spectral bands. Also expect network downloads, local cached data, and non-reproducible dependency resolution unless versions are pinned.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (10)

Tp4

High

• Category: MCP Tool Poisoning
• Confidence: 91% confidence
• Finding: The skill is presented as a local NDVI before/after analyzer, but the documentation also introduces automatic remote data acquisition and alternate input types that materially expand its behavior. This can mislead operators about network access, provenance of analyzed data, and what asset types are accepted, increasing the chance of unsafe or incorrect processing in compliance workflows.

Description-Behavior Mismatch

Medium

• Confidence: 87% confidence
• Finding: The documentation adds an automatic remote download workflow from Planetary Computer, expanding the trust boundary from local files to external data sources. In a compliance-oriented skill, silent network retrieval can introduce data provenance, privacy, reproducibility, and unexpected egress risks if users believe the tool only analyzes supplied rasters.

Intent-Code Divergence

Medium

• Confidence: 84% confidence
• Finding: The CLI docs conflict about whether the tool expects NDVI rasters, raw images, or bbox/date-driven downloads, which can cause operators to provide unintended inputs or trigger hidden behaviors. In an enforcement or compliance context, ambiguous input semantics can lead to incorrect outputs and make it harder to assess whether the skill is accessing data or processing content as expected.

Intent-Code Divergence

High

• Confidence: 96% confidence
• Finding: The auto-download helper explicitly says it downloads Sentinel-2 visual previews and assigns them to args.before_ndvi and args.after_ndvi, causing the downstream pipeline to treat non-NDVI imagery as if it were NDVI rasters. In this skill's compliance context, that can produce materially false land-use change findings and compliance reports, creating a high-integrity failure that could trigger wrongful investigations or decisions.

Description-Behavior Mismatch

High

• Confidence: 98% confidence
• Finding: The auto-download path advertises NDVI-based cropland change analysis but prefers assets such as visual and thumbnail before spectral bands, then passes the downloaded files directly into compute_change as if they were NDVI rasters. This is a concrete data/semantic mismatch that can silently corrupt analysis results in a regulatory/compliance workflow, making the skill especially dangerous because users may trust the output as evidence.

Unpinned Dependencies

Low

• Category: Supply Chain

Content

geoskill-data-fetcher>=0.2.0

# Third-party dependencies used by this skill
numpy
rasterio
shapely

• Confidence: 94% confidence
• Finding: numpy

Unpinned Dependencies

Low

• Category: Supply Chain

Content

# Third-party dependencies used by this skill
numpy
rasterio
shapely

• Confidence: 94% confidence
• Finding: rasterio

Unpinned Dependencies

Low

• Category: Supply Chain

Content

# Third-party dependencies used by this skill
numpy
rasterio
shapely

• Confidence: 93% confidence
• Finding: shapely

Unpinned Dependencies

Low

• Category: Supply Chain

Content

# Requirements for cropland-change-compliance
# Shared data-download library (Planetary Computer / NASA POWER / OSM)
geoskill-data-fetcher>=0.2.0

# Third-party dependencies used by this skill
numpy

• Confidence: 90% confidence
• Finding: geoskill-data-fetcher>=0.2.0

Known Vulnerable Dependency: numpy — 10 advisory(ies): CVE-2014-1859 (Numpy arbitrary file write via symlink attack); CVE-2021-41495 (NumPy NULL Pointer Dereference); CVE-2021-33430 (NumPy Buffer Overflow (Disputed)) +7 more

Critical

• Category: Supply Chain
• Confidence: 72% confidence
• Finding: numpy

VirusTotal

VirusTotal findings are pending for this skill version.

Static analysis

No suspicious patterns detected.