skill-creator

Warn

Audited by Gen Agent Trust Hub on Aug 31, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPRIVILEGE_ESCALATIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill provides instructions and commands for fetching agent skills from external URLs or upstream repository references provided by the user.\n- [REMOTE_CODE_EXECUTION]: The skill guides the agent to use a bootstrap script that downloads and applies external code to the local environment, representing a risk if the source is not verified.\n- [COMMAND_EXECUTION]: The skill relies on various shell and Python scripts to perform local file system operations, registry updates, and repository maintenance tasks.\n- [PRIVILEGE_ESCALATION]: The skill initialization utility automatically grants executable permissions (chmod 0o755) to generated scripts.\n- [INDIRECT_PROMPT_INJECTION]: The skill ingests and processes external skill data, creating a vulnerability surface where malicious instructions could be introduced.\n
  • Ingestion points: External URLs and repository references used during the skill import process.\n
  • Boundary markers: No specific markers or delimiters are defined to isolate or warn against instructions within external content.\n
  • Capability inventory: The skill possesses extensive capabilities for file modification, script execution, and permission management.\n
  • Sanitization: The provided validation script focuses on structural and YAML formatting checks but does not perform security-focused sanitization of the content.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Aug 31, 2026, 05:40 AM
Security Audit — agent-trust-hub — skill-creator