Skill Preflight Bootstrap

• Outcome: Pass
• Latest audit: May 29, 2026 at 8:15 PM UTC
• Version:
  0.1.0

Overview

The skill transparently bootstraps a repository-level skill-preflight workflow, with scoped project file changes and no evidence of hidden data theft or destructive behavior.

Use `--dry-run` first and review the generated `.codex/settings.json`, `.claude/settings.json`, `AGENTS.md`, and hook scripts before applying. Be aware that remote skill search runs an external `npx` command and may send the task description to that tool, so avoid including secrets in those queries or in `.learnings` notes.

SkillSpector

By NVIDIA

Vulnerability Patterns

• MCP Least Privilege: Underdeclared Capability, Wildcard Permission, Missing Permission Declaration
• 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 (1)

Lp3

Medium

• Category: MCP Least Privilege
• Confidence: 93% confidence
• Finding: The skill instructs the operator to run a bootstrap script that reads from the current skill directory and writes multiple files into a target repository, including scripts, hooks, settings, and AGENTS.md content. Those are code-capable behaviors, but the skill declares no permissions, which weakens reviewability and can cause users or enforcement systems to authorize powerful file and shell operations without explicit disclosure.

VirusTotal

65/65 vendors flagged this skill as clean.