skill-installer

Pass

Audited by Gen Agent Trust Hub on Jun 18, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: Downloads data from GitHub API and download servers. Implements path validation during ZIP extraction in install-skill-from-github.py to prevent directory traversal.
  • [COMMAND_EXECUTION]: Uses git via list-based subprocess calls in install-skill-from-github.py, mitigating shell injection risks.
  • [PROMPT_INJECTION]: The skill handles untrusted external data which presents an indirect prompt injection surface.
  • Ingestion points: list-skills.py (remote directory names) and install-skill-from-github.py (remote file content).
  • Boundary markers: Absent in instructions.
  • Capability inventory: subprocess.run, file-write, and network operations are available across scripts.
  • Sanitization: Validates relative paths and skill names to prevent file system abuse.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 18, 2026, 02:45 AM
Security Audit — agent-trust-hub — skill-installer