skills/openai/skills/skill-installer/Gen Agent Trust Hub

skill-installer

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONREMOTE_CODE_EXECUTION
Full Analysis
  • External Content Downloads: The skill retrieves repository contents and archives from GitHub domains (github.com and codeload.github.com). This mechanism is necessary for the skill's primary function of installing new capabilities from remote sources.
  • System Command Invocation: The skill uses subprocess.run to execute the git CLI for cloning and checking out specific repository paths. The implementation mitigates common command injection risks by passing command arguments as a structured list rather than a single string.
  • Archive Security Validation: A specialized check in the extraction process (_safe_extract_zip) ensures that all files within a downloaded archive are placed within the designated destination. This is a standard security practice that helps prevent directory traversal (ZipSlip) vulnerabilities.
  • Credential Integration: The utility can utilize environment variables like GITHUB_TOKEN to access private repositories or avoid API rate limits. This follows established secret management patterns by relying on the execution environment's configuration rather than hardcoding credentials.
  • Path and Name Validation: The skill includes explicit validation for skill names and paths to ensure they are relative and do not attempt to access or overwrite locations outside of the intended directory structure.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 12:14 AM
Security Audit — agent-trust-hub — skill-installer