ClawHub
Back to skill

Security audit

Git Reporter

Security checks across malware telemetry and agentic risk

Overview

This skill locally reads git repository metadata to draft standups and reports, and I found no hidden network access, persistence, or destructive behavior.

Install only if you are comfortable with the agent reading the current repository's git metadata and working-tree summaries. Review generated reports before sharing them, especially if commit messages, branch names, stash labels, or file paths may reveal sensitive project details.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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
90% confidence
Finding
The natural-language triggers are very broad everyday phrases like '帮我写今天的站会' and '生成一下本周的周报', which can cause the skill to activate in contexts where the user did not explicitly intend to invoke repository analysis. Because this skill reads local git metadata, commit history, branch names, stash entries, and diffs, accidental activation can expose sensitive local project information in generated output or to downstream model context.

Natural-Language Policy Violations

Medium
Confidence
80% confidence
Finding
Automatically switching output language based on commit-message language changes behavior using repository contents rather than explicit user choice. Since commit messages are untrusted input and may be mixed-language or attacker-influenced, this can produce confusing, manipulated, or policy-bypassing output and reduces user control over how repository-derived content is presented.

Vague Triggers

Medium
Confidence
88% confidence
Finding
The auto-trigger description is broad enough to match common phrases like asking for a daily report or weekly summary, which can invoke the skill unexpectedly in unrelated contexts. Because the skill runs local git commands and summarizes repository state, unintended activation can expose commit history, branch names, author identity, and uncommitted work to the conversation without sufficiently explicit user intent.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.