ClawHub
Back to skill

Security audit

commit-messages

Security checks across malware telemetry and agentic risk

Overview

This skill is a small commit-message helper that inspects staged git changes and writes a local draft message, with some trigger and file-write caveats but no hidden or destructive behavior.

Install only if you are comfortable with the skill reading staged diffs and recent git history when invoked. Watch for accidental activation from broad git or commit wording, and check whether ./commit_msg.txt already exists before using it in a repository.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Vague Triggers

Medium
Confidence
91% confidence
Finding
The trigger list includes broad terms like "git" and especially "commit", which can cause this skill to activate in many routine contexts beyond the narrow task of drafting commit messages. Over-broad auto-invocation increases the chance the agent will inspect repository state and staged diffs unexpectedly, potentially exposing sensitive code or metadata in situations where the user did not intend to invoke this skill.

Missing User Warnings

Medium
Confidence
88% confidence
Finding
The skill instructs the agent to write output directly to ./commit_msg.txt without explicitly warning the user or requesting confirmation. Silent file creation/modification is risky because it changes the working directory state and may overwrite files or leave artifacts the user did not expect.

Shadow Command Trigger

Medium
Category
Trigger Abuse
Confidence
95% confidence
Finding
Using the trigger term "commit" conflicts with a built-in command name and creates a shadow-command risk. This can route a user asking to perform a normal commit operation into this skill instead, causing unintended behavior such as repository inspection and file writes rather than the expected built-in flow.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.