composition-quality-checker

Pass

Audited by Gen Agent Trust Hub on Nov 7, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill does not contain any malicious patterns such as prompt injection, obfuscation, or data exfiltration. Its scope is restricted to providing Pass/Warn/Fail verdicts and remediation hints based on provided input specs.
  • [COMMAND_EXECUTION]: The skill explicitly states in its official runtime boundary and non-goals that it does not execute Python, run Blender, or move cameras. It only validates data snapshots.
  • [DATA_EXFILTRATION]: There are no network-capable commands (e.g., curl, wget, fetch) or references to external, non-whitelisted domains. All analysis is performed locally within the agent's context.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input such as 'viewport_or_render_preview_notes'. However, the threat is negligible because the skill possesses no exploitable capabilities like file system access, network communication, or shell execution to act upon malicious instructions embedded in that data.
  • [REMOTE_CODE_EXECUTION]: No external dependencies or remote scripts are downloaded or executed. The skill relies entirely on internal markdown reference files for its logic.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 7, 2026, 09:09 PM