defold-project-setup

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill fetches engine version metadata from d.defold.com and downloads engine releases from the official Defold organization repository on GitHub.
  • [INDIRECT_PROMPT_INJECTION]: The script reads and processes external dependency URLs defined within the local game.project file.
  • Ingestion points: Reads dependency URLs and library include directories from game.project in the project root.
  • Boundary markers: Absent; the script assumes URLs in the project configuration are valid targets for download.
  • Capability inventory: Performs network requests (urllib.request), directory deletion (shutil.rmtree), and file extraction (zipfile).
  • Sanitization: Implements strict validation using assert_safe_include_dir (regex-based whitelist) and enforces path boundaries during extraction to mitigate traversal risks.
  • [COMMAND_EXECUTION]: Instructions provide a direct path to execute a local Python script (fetch_deps.py) to perform the project synchronization.
  • [SAFE]: The extraction logic includes a security check that resolves output paths and verifies they reside within the designated .deps/ folder, effectively preventing zip-slip attacks.
  • [SAFE]: Dependency directory names are validated against a restricted character set ([A-Za-z0-9_-]+) and checked for path traversal sequences like .. or slashes before processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 05:23 PM
Security Audit — agent-trust-hub — defold-project-setup