Score a produced visual asset (image / graphic / carousel frame) against its brief — render fidelity + brand-fit — inside an existing eval loop, using the re-ingested asset (the return-leg). One asset (or picked variant) per cycle, verdict + brief-fidelity diagnosis. Not for video (use evaluate-shortform), a landing page (use evaluate-landing-page), live-post engagement (use evaluate-content / evaluate-ad), generating the asset (use produce-asset), or scaffolding the loop (use run-pipeline).
$npx -y skills add hungv47/meta-skills --skill evaluate-assetInstalls into the current project.
Run `npx skills use "https://github.com/hungv47/meta-skills" --skill "hungv47/meta-skills/evaluate-asset"` and follow the generated skill instructions now. Read its complete output, redirecting it to a temporary file first if necessary. Resolve relative paths from the supporting-files directory it provides.
Use the skills in "https://github.com/hungv47/meta-skills" that are relevant to the current task. Run `npx skills add "https://github.com/hungv47/meta-skills"` and select the relevant skills, then follow their instructions.