higgsfield-soul-id
Fail
Audited by Gen Agent Trust Hub on Jun 13, 2026
Risk Level: CRITICALREMOTE_CODE_EXECUTIONCOMMAND_EXECUTION
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill bootstrap instructions include a command to download and execute a shell script from
https://raw.githubusercontent.com/higgsfield-ai/cli/main/install.shby piping it directly tosh. This is a high-risk practice that allows arbitrary code execution from a third-party source without verification or user review. - [COMMAND_EXECUTION]: The skill heavily utilizes the
Bashtool to interact with the Higgsfield CLI. It interpolates user-supplied arguments, such as character names and file paths, directly into shell commands (e.g.,higgsfield soul-id create --name "<name>" ...). Without explicit sanitization or strict boundary markers, this pattern is susceptible to command injection attacks. - [INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface as it ingests untrusted data (user-provided names and photo paths) and uses them in sensitive capabilities (bash command execution). No sanitization or boundary markers are employed to prevent malicious input from manipulating the intended shell command behavior.
Recommendations
- HIGH: Downloads and executes remote code from: https://raw.githubusercontent.com/higgsfield-ai/cli/main/install.sh - DO NOT USE without thorough review
- AI detected serious security threats
Audit Metadata