anti-reversing-techniques

Pass

Audited by Gen Agent Trust Hub on Feb 27, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONREMOTE_CODE_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: Indirect Prompt Injection Surface.
  • Ingestion points: The skill processes untrusted external data (binaries and malware) as described in resources/implementation-playbook.md.
  • Boundary markers: None identified. There are no instructions to delimit or ignore instructions that might be embedded in the data being analyzed.
  • Capability inventory: The skill provides methods for binary patching (ida_bytes.patch_byte), library hooking (LD_PRELOAD), and process tracing (ptrace).
  • Sanitization: None mentioned for the data processed.
  • [COMMAND_EXECUTION]: The playbook includes specific shell commands for environment manipulation and binary decompression, such as 'LD_PRELOAD=./hook.so ./target' and 'upx -d packed.exe'.
  • [REMOTE_CODE_EXECUTION]: The skill provides scripts for dynamic code modification and execution flow redirection, including C code for hooking system calls and IDAPython snippets for patching binary checks.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 27, 2026, 09:00 AM