ClawHub
Back to skill

Security audit

ia-receiving-code-review

Security checks across malware telemetry and agentic risk

Overview

This skill is a coherent code-review workflow that can read and reply to PR comments, which fits its stated purpose but should be used with normal GitHub credential care.

Install this if you want an agent to critically triage and respond to code review feedback. Use it in repositories where the agent's GitHub credentials are allowed to read PR comments and post replies, and review proposed public-facing replies when the discussion is sensitive or high impact.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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 (2)

Context-Inappropriate Capability

Medium
Confidence
88% confidence
Finding
The skill explicitly instructs querying GitHub PR comments via `gh api`, which expands its behavior from passive review analysis into live network access against a repository. Even if read-only in this instance, it creates an external-capability dependency and broadens the attack surface because untrusted PR context can drive API calls and data retrieval beyond the narrowly stated purpose of critically evaluating feedback.

Context-Inappropriate Capability

Medium
Confidence
95% confidence
Finding
This instruction tells the agent to post inline PR replies using the GitHub API, which is a remote write action not implied by the skill's description of critically processing review feedback. Remote writes based on potentially ambiguous or adversarial review content can lead to unauthorized repository interaction, spam, disclosure, or workflow manipulation if the skill is triggered automatically or with broad credentials.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.