The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/agent.py is vulnerable to shell command injection. The run_vol3 function uses subprocess.run(cmd, shell=True) where the cmd is constructed using f-strings that incorporate variables like memory_dump and extra_args directly from input. Without strict validation, an attacker could supply a malicious filename or argument (e.g., ; rm -rf /) to execute arbitrary commands on the host system. This risk is further present in the scan_with_yara and detect_code_injection functions which pass unvalidated string arguments to the shell execution helper.
[CREDENTIALS_UNSAFE]: The skill provides automated capabilities for harvesting sensitive system secrets. The dump_credentials function in scripts/agent.py and the instructions in SKILL.md (Step 5) specifically guide the agent to use Volatility plugins (windows.hashdump, windows.cachedump, and windows.lsadump) to extract Windows SAM hashes, domain cached credentials, and LSA secrets. While this is a standard forensic procedure, it constitutes a high-risk capability for an AI agent that could lead to unauthorized credential exposure.

analyzing-memory-dumps-with-volatility