debug-game-in-emulator
Pass
Audited by Gen Agent Trust Hub on Oct 6, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands to launch emulator binaries, verify software versions, and attach system debuggers such as lldb and gdb. These actions are fundamental to the skill's stated purpose of software debugging.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from external log files and command output, which presents a surface for indirect prompt injection.
- Ingestion points: The agent is required to read emulator-generated logs (e.g., duckstation.log, RPCS3.log, emulog.txt) and binary help/version output.
- Boundary markers: The skill contains instructions to "Quote a log line" and "Classify the failure by layer", which encourages selective processing, but it lacks formal delimiters or specific warnings to ignore instructions that might be embedded in the ingested logs.
- Capability inventory: The agent has the ability to execute shell commands to launch processes, modify configuration files, and attach debuggers to running processes.
- Sanitization: The skill does not define specific sanitization or filtering logic for external text, although it mentions stripping ANSI colors from console output.
Audit Metadata