Security audit

data-parser-toolkit

Security checks across malware telemetry and agentic risk

Overview

This is a normal data-file parsing skill, but users should be careful with remote URL parsing and unpinned parser dependencies.

Review and pin dependency versions before using this in production or on sensitive data. Avoid or restrict parse_from_url in networks where internal services are reachable, and sandbox parsing of untrusted spreadsheets or Parquet files. Treat the README's unrelated import examples as documentation noise, not required setup.

SkillSpector

By NVIDIA

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands

Findings (10)

Intent-Code Divergence

Medium

Confidence: 89% confidence
Finding: The README claims integration into an OpenClaw/Claude Code skill system and shows imports of unrelated modules not otherwise described by this toolkit. In an agent-skill context, misleading integration guidance can cause users or automated systems to trust and load unexpected local modules from a skills path, increasing the risk of dependency confusion, unintended code execution, or use of components outside the reviewed scope.

Context-Inappropriate Capability

Medium

Confidence: 82% confidence
Finding: The skill includes arbitrary URL fetching and remote content parsing, which introduces network egress and the possibility of SSRF-style access to internal services or unintended retrieval of sensitive resources if attacker-controlled URLs are passed in. In a data-parsing skill, remote fetch capability is more dangerous because users may expect local file parsing only, while the code can contact external or internal endpoints silently.

Missing User Warnings

Medium

Confidence: 83% confidence
Finding: The README advertises parsing from remote URLs and retry behavior without any warning about trust boundaries, data integrity, authentication, or privacy implications. In a skill/agent setting, encouraging remote fetches can lead to SSRF-like behavior, ingestion of malicious content, or accidental exfiltration through requests to attacker-controlled endpoints.

Missing User Warnings

Medium

Confidence: 85% confidence
Finding: `read_from_url()` performs outbound network access without any user-facing notice, confirmation, or policy guardrails. That creates privacy and SSRF risk because a caller may unintentionally cause the skill to contact attacker-chosen endpoints, including internal services, while processing seemingly ordinary input.

Unpinned Dependencies

Low

Category: Supply Chain
Content: pandas openpyxl chardet pyarrow
Confidence: 95% confidence
Finding: The dependency 'pandas' is unpinned, so installs may resolve to different versions over time, including versions with newly introduced vulnerabilities or breaking changes. This weakens supply-chain integrity and makes builds non-reproducible, which is a genuine security concern even though it is not immediately exploitable on its own.

Unpinned Dependencies

Low

Category: Supply Chain
Content: pandas openpyxl chardet pyarrow xlrd>=2.0.0
Confidence: 95% confidence
Finding: The dependency 'openpyxl' is unpinned, allowing package resolution to drift between installations. This can silently introduce vulnerable or incompatible versions and reduces reproducibility of the environment.

Unpinned Dependencies

Low

Category: Supply Chain
Content: pandas openpyxl chardet pyarrow xlrd>=2.0.0
Confidence: 95% confidence
Finding: The dependency 'chardet' is unpinned, which permits uncontrolled upgrades or resolver-selected versions at install time. That creates avoidable supply-chain risk and makes security review of the environment unreliable.

Unpinned Dependencies

Low

Category: Supply Chain
Content: pandas openpyxl chardet pyarrow xlrd>=2.0.0
Confidence: 95% confidence
Finding: The dependency 'pyarrow' is unpinned, so future installs may fetch different versions, including ones with known vulnerabilities. In a package that processes structured data files, this increases the chance of pulling in a risky version without notice.

Known Vulnerable Dependency: openpyxl — 2 advisory(ies): CVE-2017-5992 (Improper Restriction of XML External Entity Reference in Openpyxl); CVE-2017-5992 (Openpyxl 2.4.1 resolves external entities by default, which allows remote attack)

High

Category: Supply Chain
Confidence: 81% confidence
Finding: openpyxl has historical XXE-related advisories, and this requirements file leaves the package unpinned, which could allow installation of an affected version. In a skill likely handling spreadsheet files, parsing attacker-controlled documents with a vulnerable XML parser could enable file disclosure, SSRF, or related parser abuse.

Known Vulnerable Dependency: pyarrow — 8 advisory(ies): CVE-2023-47248 (PyArrow: Arbitrary code execution when loading a malicious data file); CVE-2019-12408 (Missing Initialization of Resource in Apache Arrow); CVE-2019-12410 (Missing Initialization of Resource in Apache Arrow) +5 more

Critical

Category: Supply Chain
Confidence: 95% confidence
Finding: pyarrow has multiple serious advisories, including arbitrary code execution when loading malicious data files, and the dependency is unpinned in a data-processing context. Because this skill appears oriented around tabular file handling, accepting untrusted Arrow/Parquet or related data formats could make exploitation materially more plausible and severe.

VirusTotal

67/67 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.