blucli

Warn

Audited by Gen Agent Trust Hub on Jul 9, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill's metadata includes an installation routine that downloads and compiles the blucli utility from an external GitHub repository (github.com/steipete/blucli) using the Go package manager.
  • [COMMAND_EXECUTION]: The skill is centered around executing the blu CLI tool to perform hardware discovery and control audio playback on the local network.
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by interpolating user-supplied strings into shell commands.
  • Ingestion points: User-provided device identifiers (IDs, names, or aliases) and TuneIn search/play queries specified in SKILL.md.
  • Boundary markers: The instructions lack delimiters or warnings to treat embedded user data as non-executable content.
  • Capability inventory: The skill invokes the blu binary through shell commands to interact with network hardware.
  • Sanitization: No input validation, escaping, or sanitization logic is present to mitigate command injection risks from adversarial input.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Jul 9, 2026, 10:44 PM
Security Audit — agent-trust-hub — blucli