ai-voice-cloning
Fail
Audited by Gen Agent Trust Hub on Feb 19, 2026
Risk Level: HIGHREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- REMOTE_CODE_EXECUTION (HIGH): The skill instructs the agent to execute a remote shell script via
curl -fsSL https://cli.inference.sh | sh. This is a classic 'curl pipe to sh' attack vector that executes unverified code from an untrusted domain. - EXTERNAL_DOWNLOADS (MEDIUM): The skill references several external packages via
npx skills add inference-sh/skills@.... These dependencies are hosted on an untrusted third-party repository and are executed at runtime. - COMMAND_EXECUTION (LOW): The skill defines
Bash(infsh *)as an allowed tool, granting the agent the ability to execute any command within theinfshCLI namespace, which is installed via the untrusted script above. - INDIRECT_PROMPT_INJECTION (LOW): The skill processes user-supplied text through the
inputfield of several models. This represents an attack surface where malicious input could influence the agent's behavior if not properly sanitized. - Ingestion points: The
textparameter in theinfsh app runcommands (SKILL.md). - Boundary markers: None present in the provided examples.
- Capability inventory: The skill allows arbitrary Bash execution for the
infshtool. - Sanitization: No evidence of sanitization or validation of the input text before processing.
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