unity-play-harness

Pass

Audited by Gen Agent Trust Hub on Nov 6, 2026

Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses the Python subprocess.run module to execute local commands for Unity Editor control and self-testing.
  • Evidence: In scripts/probe.py, the run_eval function invokes the unity command-line tool.
  • Evidence: In scripts/journey_report.py, the selftest function executes the script itself using the Python interpreter.
  • [DYNAMIC_EXECUTION]: The harness facilitates the execution of dynamic C# code snippets within the Unity Editor process.
  • Evidence: scripts/probe.py evaluates strings such as return GameProbe.Json(); via the Unity CLI's eval command.
  • Evidence: resources/Editor/ProbeCommands.cs registers Editor-side commands that execute logic on the main thread during Play mode.
  • [INDIRECT_PROMPT_INJECTION]: The scripts/journey_report.py script presents a surface for indirect prompt injection by processing external telemetry data.
  • Ingestion points: scripts/journey_report.py reads and parses timeline.jsonl files from the local file system.
  • Boundary markers: There are no delimiters or explicit instructions to the AI to ignore embedded commands within the log data.
  • Capability inventory: The skill can execute Unity commands and write files based on the interpreted results of these reports.
  • Sanitization: No semantic validation or sanitization is performed on the ingested game state values before they are incorporated into the report output.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 6, 2026, 06:19 AM