facenet-cli
Pass
Audited by Gen Agent Trust Hub on Mar 29, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructs the agent to install a software package directly from a remote Git repository (https://github.com/mz038197/facenet-cli.git) using the uv package manager. This repository belongs to the skill author and is the primary source for the tool.
- [REMOTE_CODE_EXECUTION]: By recommending the installation and execution of the facenet-cli tool from GitHub, the skill enables the execution of remote code. This is necessary for the skill's stated purpose of performing facial recognition.
- [COMMAND_EXECUTION]: The skill uses shell commands to operate the fnet CLI, including subcommands for camera capture (embedding capture), batch image processing (embedding images), and running tests (pytest).
- [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection as it processes external data (images and CSV databases) and returns the results as JSON for agent parsing. Ingestion points: The tool reads data from user-provided image files (--image1, --image2) and CSV databases (--database-csv). Boundary markers: The instructions lack explicit boundary markers or warnings for the agent to ignore instructions embedded in file metadata or CSV content. Capability inventory: The skill can execute shell commands and write to the local file system. Sanitization: There is no evidence of sanitization or filtering of external data before it is structured into JSON for the agent.
Audit Metadata