ClawHub

Repo Guardian

Security checks across malware telemetry and agentic risk

Overview

Repo Guardian is transparent about reviewing GitHub pull requests, but it can automatically merge code using AI approvals and a write-capable GitHub token, so it needs careful review before use.

Install only if you intend to delegate PR review and possible merges to automation. Start with --dry-run or GUARDIAN_AUTO_MERGE=false, use a fine-grained GitHub token limited to the target repository, configure a genuinely separate Reviewer B agent, and rely on GitHub branch protection and CI checks as hard safeguards before allowing unattended cron runs.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Missing User Warnings

Medium
Confidence
90% confidence
Finding
The script sends full PR titles, branch names, file lists, and diffs to external agent processes for analysis without a strong explicit consent gate or data-classification check. In the context of a repository-governance skill, this can expose proprietary code, credentials accidentally present in diffs, or sensitive metadata to downstream model providers or local integrations.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The script automatically merges pull requests whenever both model outputs say APPROVE, without a human confirmation step, branch-protection verification, or robust validation of review quality. In a governance automation skill, this is risky because prompt-injected PR content or weak/failed model reviews could directly trigger unauthorized or unsafe code changes into the repository.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal