adaptive-glass
Pass
Audited by Gen Agent Trust Hub on Nov 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses the 'ffprobe' utility to extract duration and resolution metadata from video files. This execution is performed using 'child_process.spawn' with a discrete argument array, which effectively prevents shell injection vulnerabilities.
- [DYNAMIC_EXECUTION]: The skill generates an 'index.html' file that includes dynamically constructed JavaScript timelines for animations. The generation logic uses regular expression validation for identifiers and performs HTML entity escaping for user-supplied strings, mitigating the risk of script injection in the generated output.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted configuration data from 'scenes.json' files. It maintains a secure processing chain by validating scene IDs against a strict allowlist and sanitizing all text fields before they are incorporated into the visual composition.
Audit Metadata