byted-mediakit-shared

Pass

Audited by Gen Agent Trust Hub on Nov 5, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data which could contain malicious instructions.
  • Ingestion points: Processes external media via --video-url, --sub-image-url, and subtitle data via --subtitles or @file pointers as seen in SKILL.md.
  • Boundary markers: No explicit instruction-ignoring delimiters or safety boundaries are defined for the content of the fetched media or subtitle files.
  • Capability inventory: The skill has shell permissions and executes commands via mediakit-cli to perform file operations, network requests for media fetching, and complex media processing as documented in SKILL.md.
  • Sanitization: The instructions provide specific shell escaping and JSON formatting guidelines for parameters, but do not specify sanitization for the content of the external files themselves.
  • [EXTERNAL_DOWNLOADS]: The skill requires downloading and installing a CLI tool at runtime.
  • Evidence: SKILL.md instructs the user to run npx @volcengine/mediakit-cli install -y for initial setup and mediakit-cli update for maintenance. These downloads originate from the verified vendor's own package registry.
  • [COMMAND_EXECUTION]: The skill relies extensively on executing shell commands to function.
  • Evidence: The skill documentation in SKILL.md and reference/query_task.md centers on constructing and executing mediakit-cli commands with various parameters, including environment variable injection for authentication and runtime context.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 5, 2026, 09:02 AM