mini-swe-agent

• Outcome: Review
• Latest audit: May 29, 2026 at 9:22 PM UTC
• Version:
  1.0.0

Overview

This skill is not malicious, but it directs an agent to run an external coding tool in autonomous no-confirmation mode that can inspect and edit a repository.

Install only if you intentionally want an external autonomous coding agent to make repository changes without stopping for approval. Verify the `mini` CLI source and version, run it in a disposable branch or sandbox, and review diffs and tests before keeping changes.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (2)

Missing User Warnings

Medium

• Confidence: 94% confidence
• Finding: The skill explicitly instructs use of `mini --yolo`, which delegates end-to-end code exploration and editing to an autonomous subprocess without human confirmation gates. In a security-sensitive agent environment, this increases the chance of unintended code changes, unsafe command execution, or modification of the wrong files because the operator is not warned about the autonomy and edit scope.

Missing User Warnings

Low

• Confidence: 88% confidence
• Finding: The installation instructions provide a shell command that creates a directory and writes a skill file into the user's persistent `~/.openclaw/skills` location without any warning about filesystem modification. While common for installation steps, silently directing users to write agent-executable content into a trusted skills directory can normalize persistent changes and reduce scrutiny of what is being installed.

VirusTotal

64/64 vendors flagged this skill as clean.