shellcode-analyzer

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill reads raw shellcode as a hex string and instructs the agent to emit that exact hex (e.g., printing it via the python command and passing it verbatim as hex_data to mcp__x64dbg__write_memory), which requires the LLM to handle and output the exact secret/binary bytes.

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (high risk: 0.90). The skill directs the agent to load, write, and execute arbitrary native shellcode inside a debugger (allocating/writing process memory, setting registers, running code), which can run destructive or networked payloads on the host and thus risks compromising machine state even without explicit sudo or system-file modification.