ai-music-generation
Fail
Audited by Gen Agent Trust Hub on Feb 18, 2026
Risk Level: HIGHREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- Remote Code Execution (CRITICAL): The skill documentation instructs users to execute 'curl -fsSL https://cli.inference.sh | sh', which downloads and executes a script from an untrusted remote source directly in the shell. This pattern provides no opportunity for the user or agent to verify the code before execution.
- External Downloads (HIGH): The skill relies on software from 'inference.sh' and utilizes 'npx' to pull external skill packages. Neither the domain nor the package author are on the trusted list of organizations or repositories.
- Command Execution (MEDIUM): The 'allowed-tools' metadata grants the agent broad permission to run any command starting with 'infsh', which significantly increases the risk if the third-party tool has vulnerabilities or hidden functionalities.
- Indirect Prompt Injection (LOW): The skill is susceptible to indirect prompt injection because it interpolates user-controlled strings into command-line arguments. 1. Ingestion points: The 'prompt' and 'lyrics' fields in the input JSON. 2. Boundary markers: Absent; there are no delimiters or instructions to ignore embedded commands. 3. Capability inventory: The skill has Bash execution capabilities via the 'infsh' tool. 4. Sanitization: Absent; user input is passed directly to the CLI tool.
Recommendations
- HIGH: Downloads and executes remote code from: https://cli.inference.sh - DO NOT USE without thorough review
- AI detected serious security threats
Audit Metadata