background-removal

Pass

Audited by Gen Agent Trust Hub on Jul 16, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill manages dependencies by fetching and installing the 'belt' CLI and related skill packages ('belt-sh/cli', 'inference-sh/skills') from the inference.sh ecosystem using the 'npx skills add' command.
  • [COMMAND_EXECUTION]: Utilizes the Bash tool to execute commands through the 'belt' CLI utility, allowing the agent to run remote AI applications like 'infsh/birefnet' and 'falai/reve'.
  • [PROMPT_INJECTION]: Detects a surface for indirect prompt injection (Category 8) where user-controlled strings such as image URLs and editing prompts are passed to remote inference applications.
  • Ingestion points: The 'image_url' and 'prompt' parameters provided in JSON inputs to 'belt app run' within 'SKILL.md'.
  • Boundary markers: Absent; user inputs are interpolated directly into JSON strings used as CLI arguments.
  • Capability inventory: The skill is granted 'Bash' tool access specifically to interact with the 'belt' executable.
  • Sanitization: No explicit input validation or sanitization logic is implemented within the skill instructions, relying on the platform's backend to handle potentially malicious payloads.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 16, 2026, 04:51 PM
Security Audit — agent-trust-hub — background-removal