Security audit

Hugo Blog Tool Free

Security checks across malware telemetry and agentic risk

Overview

This Hugo publishing skill is mostly purpose-aligned, but it gives an agent broad publish, local-reading, and credential-setup guidance without enough user confirmation or scoping.

Install only if you are comfortable with an agent editing your Hugo repository, committing, and pushing to a remote. Before use, require confirmation for the target blog path, files changed, commit message, remote branch, and any credential or SSH-key setup; do not allow automatic reading of memory files or global Git config unless you explicitly approve it.

SkillSpector

By NVIDIA

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands

Findings (8)

Context-Inappropriate Capability

Medium

Confidence: 91% confidence
Finding: The skill advises setting a global Git credential helper cache, which persists authentication behavior beyond the narrow task of publishing a single Hugo post. This expands the agent's scope into long-lived credential management and can weaken least-privilege expectations or affect unrelated repositories and workflows on the host.

Context-Inappropriate Capability

Medium

Confidence: 88% confidence
Finding: Including SSH key generation as part of the skill extends it from content publishing into account and authentication setup. Key generation and account binding are sensitive security operations that can create persistent access paths and should not be bundled into a routine blog-post publishing workflow.

Intent-Code Divergence

Low

Confidence: 80% confidence
Finding: The trigger-condition text says the skill should be used for marketing copy and content creation, which conflicts with the rest of the document describing a Hugo publishing tool. This mismatch can cause accidental invocation in unrelated contexts, increasing the chance that file-modifying or git-pushing behavior is activated unexpectedly.

Vague Triggers

High

Confidence: 85% confidence
Finding: The capability and triggering language is broad enough that the skill may activate on ordinary writing or blog-help prompts rather than explicit publish requests. Because this skill can write files and push to a remote repository, overbroad activation materially raises the risk of unintended side effects from casual user interactions.

Vague Triggers

High

Confidence: 92% confidence
Finding: The documented trigger scope is inconsistent with the real behavior of the skill and lacks clear boundaries. In a skill that can edit local content and push to remotes, ambiguous activation criteria make accidental execution more dangerous than in a read-only or advisory-only tool.

Missing User Warnings

Medium

Confidence: 94% confidence
Finding: The skill describes automatic file modification and remote git push behavior without an explicit warning or confirmation step. In context, this is risky because publishing content and altering a repository are irreversible or externally visible actions that should require informed user approval.

Missing User Warnings

Medium

Confidence: 90% confidence
Finding: The skill says it will automatically read user memory files and local Git configuration before asking the user, but does not warn about privacy implications or seek consent. This can cause the agent to ingest and reuse sensitive local configuration or stored personal data in ways the user did not explicitly authorize.

Ssd 3

Medium

Confidence: 95% confidence
Finding: Automatically reading memory files and Git configuration can expose repository URLs, usernames, emails, paths, and other sensitive context without explicit consent. In this skill, that data may influence file paths, remotes, or outputs, creating a concrete privacy and unintended-data-use risk.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.