meshoptimizer

Pass

Audited by Gen Agent Trust Hub on Nov 6, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes external mesh assets (glTF, PLY, OBJ) which could potentially contain malicious payloads designed to influence agent behavior.
  • Ingestion points: SKILL.md Section 3 details processing asset files through gltfpack and meshopt API calls.
  • Boundary markers: SKILL.md Section 1 and Section 3 establish boundaries by requiring users to freeze input artifacts, define permitted output directories, and verify semantic inventories before replacements.
  • Capability inventory: The skill utilizes Read, Write, Edit, and Bash tools to interact with the filesystem and execute build or conversion utilities.
  • Sanitization: The scripts/preflight.py script validates file paths and explicitly blocks access to sensitive directories and filenames (e.g., .ssh, .env, credentials), though it does not perform deep semantic validation of the mesh data contents.
  • [COMMAND_EXECUTION]: The skill includes a Python script and provides examples for executing shell commands to build the library and convert assets.
  • The scripts/preflight.py script is provided for offline integrity checking. It uses security best practices such as os.O_NOFOLLOW and dir_fd to prevent symlink-based attacks and restricts file access to specific manifest-listed items.
  • SKILL.md and references/upstream.md provide recipes for cmake, git, and gltfpack operations, which are standard for the library's intended use case and require user approval before execution.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 6, 2026, 08:52 AM