The Agent Skills Directory

[COMMAND_EXECUTION]: The Python scripts scripts/agent.py and scripts/process.py utilize the subprocess.run and subprocess.run modules to execute external command-line tools such as nmap, netexec, and bloodhound-python. This is the intended functionality for performing network reconnaissance and security testing.- [INDIRECT_PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection by processing unverified data from the network environment.
Ingestion points: Data is ingested from the network via nmap XML output and LDAP queries in scripts/agent.py and scripts/process.py.
Boundary markers: The skill lacks explicit boundary markers or instructions for the agent to ignore potentially malicious payloads embedded in hostnames or service banners.
Capability inventory: The skill has access to the local filesystem for writing reports and can execute shell commands.
Sanitization: There is minimal sanitization of the strings retrieved from network scans before they are included in generated reports or processed further, which could allow an attacker to influence the agent's behavior through malicious network metadata.

conducting-internal-network-penetration-test