meccha-chameleon-game-trainer-analysis
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONOBFUSCATION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes the Python
ctypeslibrary to interface with sensitive Windows kernel functions, includingOpenProcess,ReadProcessMemory, andWriteProcessMemory. These functions allow the skill to interact directly with the memory of other running processes, a capability that typically requires high system privileges.\n- [INDIRECT_PROMPT_INJECTION]: The skill is designed to analyze external and potentially untrusted game trainer executables and scripts, creating a surface for indirect instructions to influence the agent. \n - Ingestion points: The skill accepts external file paths for analysis through functions like
static_analysis_workflowandanalyze_suspicious_indicators.\n - Boundary markers: The instructions do not define boundary markers or safety delimiters to isolate the content of the analyzed files from the agent's instructions.\n
- Capability inventory: The skill includes functions for process attachment, memory modification, and system-level monitoring.\n
- Sanitization: The skill lacks sanitization or validation logic to filter potentially malicious instructions embedded in the data of analyzed trainers.\n- [OBFUSCATION]: The skill includes an example of obfuscation detection, specifically providing a Base64-encoded string (
V3JpdGVQcm9jZXNzTWVtb3J5) that decodes to a sensitive memory manipulation API. This is provided to educate researchers on identifying hidden functionality in malicious tools.
Audit Metadata