Skills
skills/inference-sh-3/skills/background-removal/Gen Agent Trust Hub

background-removal

Pass

Audited by Gen Agent Trust Hub on Mar 25, 2026

Risk Level: SAFEREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The skill provides instructions to install the infsh CLI by piping a script from https://cli.inference.sh directly into the shell. This is a vendor-standard installation pattern for this tool.\n- [EXTERNAL_DOWNLOADS]: Downloads binary executables and verification checksums from dist.inference.sh as part of the CLI setup process.\n- [COMMAND_EXECUTION]: Executes shell commands via the Bash tool to run image processing apps like birefnet and reve.\n- [PROMPT_INJECTION]: The skill processes external image_url data, which presents a surface for indirect prompt injection from malicious image metadata or content.\n
  • Ingestion points: The image_url field within JSON payloads passed to infsh app run in SKILL.md.\n
  • Boundary markers: The skill uses JSON formatting for input parameters, which helps separate data from executable commands.\n
  • Capability inventory: The skill is permitted to execute shell commands via the Bash tool.\n
  • Sanitization: No evidence of input validation or content filtering for the retrieved images is present in the skill instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 25, 2026, 01:02 AM