Security audit

open-ralph

Security checks across malware telemetry and agentic risk

Overview

This is an instruction-only skill for running a bounded autonomous coding loop, and its powerful code-changing behavior is disclosed and aligned with its purpose.

Install only if you intend to let a local coding agent modify the current repository. Use it on a clean branch or worktree, keep iteration limits conservative, avoid private code or secrets unless OpenCode/provider processing is acceptable, and review the full git diff before committing or merging.

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 (1)

Missing User Warnings

Medium

Confidence: 88% confidence
Finding: The skill explicitly promotes an autonomous coding loop that can repeatedly modify a repository, but it does not clearly warn users up front that execution may change files automatically. In a code-assistant context, that omission matters because users may invoke the skill expecting analysis or troubleshooting, not persistent write actions across multiple iterations, increasing the chance of unintended or hard-to-review changes.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.