install-from-remote-library

Warn

Audited by Gen Agent Trust Hub on Apr 3, 2026

Risk Level: MEDIUMREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The skill uses npx ai-agent-skills install to download and integrate external code from arbitrary GitHub repositories. This provides a mechanism for remote code execution within the agent's environment.
  • [COMMAND_EXECUTION]: The skill executes shell commands using npx to interact with remote libraries and perform system-level installations.
  • [EXTERNAL_DOWNLOADS]: Fetches and installs content from remote Git repositories. Although the workflow mandates the use of --list and --dry-run for inspection, the final installation step introduces external, unverified code into the local environment.
  • [PROMPT_INJECTION]: The skill processes output from remote repositories via the --list command. This creates a surface for indirect prompt injection where malicious repository metadata could influence the agent's subsequent installation decisions.
  • Ingestion points: Output of npx ai-agent-skills install <source> --list (SKILL.md).
  • Boundary markers: Absent; the agent is expected to interpret the list output directly.
  • Capability inventory: Execution of npx commands and installation of files (SKILL.md).
  • Sanitization: None; the skill does not specify any validation or filtering of the remote library's output.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Apr 3, 2026, 04:17 PM