pcsx2-emulator-operations

Pass

Audited by Gen Agent Trust Hub on Sep 9, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill provides instructions in references/source-builds.md for downloading the PCSX2 source code and build dependencies from the official GitHub repository (github.com/PCSX2/pcsx2). This is a standard development workflow for a well-known project.\n- [COMMAND_EXECUTION]: The unit test script scripts/test_build_command.py uses subprocess.run to execute the local Python script build_command.py for validation purposes. The execution is limited to local scripts within the skill's own package.\n- [INDIRECT_PROMPT_INJECTION]: The skill is designed to interact with and process emulator-specific guest data, which presents an attack surface for indirect prompt injection.\n
  • Ingestion points: The agent reads guest emulator logs (emulog.txt), PS2 ELF files, disc images, and .pnach patch files to perform configuration and debugging tasks.\n
  • Boundary markers: No explicit boundary markers or instruction-ignoring warnings are implemented for the ingestion of log and data content.\n
  • Capability inventory: The skill includes local Python scripts for command generation and instructions for building and running software, though it does not provide arbitrary command execution capabilities to the guest content.\n
  • Sanitization: The build_command.py script employs shlex.join to safely quote and escape shell arguments when constructing commands, mitigating potential injection risks during command generation.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 9, 2026, 12:53 AM