background-removal

Pass

Audited by Gen Agent Trust Hub on Oct 9, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill possesses a potential surface for indirect prompt injection as it is designed to ingest and process data from arbitrary remote URLs provided by the user.
  • Ingestion points: Downloads remote image files via curl based on the --image parameter in SKILL.md.
  • Boundary markers: The instructions provide a structured processing flow, but do not specify explicit delimiters or 'ignore' instructions for the data contained within downloaded files.
  • Capability inventory: The skill uses curl for network requests and has the capability to write to the local file system in the /tmp/verging-bg-removal/ directory as described in the processing flow.
  • Sanitization: No specific sanitization or validation of the downloaded content is mentioned beyond checking supported file formats (JPG, PNG, WebP).
  • [EXTERNAL_DOWNLOADS]: The skill downloads content from external, user-specified URLs and communicates with vendor infrastructure.
  • Evidence: Uses curl to fetch images from remote URLs and uploads processed data to verging.ai and associated Cloudflare R2 storage as part of the job creation flow documented in SKILL.md.
  • [COMMAND_EXECUTION]: The skill utilizes system binaries to perform network operations.
  • Evidence: Relies on the curl binary to handle API requests, image downloads, and uploads to presigned URLs, as specified in the metadata and example commands.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 9, 2026, 03:37 PM