The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill references and provides links to external software repositories and tools that are not from trusted sources, including the Coercer tool hosted at github.com/p0dalirius/Coercer.
[COMMAND_EXECUTION]: The skill provides extensive command-line templates for executing remote commands on target systems via ntlmrelayx.py, specifically using the -c flag for SMB execution and xp_cmdshell for MSSQL targets.
[PRIVILEGE_ESCALATION]: Detailed playbooks are provided for Active Directory privilege escalation through NTLM relay to LDAP (facilitating Resource-Based Constrained Delegation and shadow credentials) and relay to ADCS HTTP endpoints for certificate enrollment (ESC8).
[DATA_EXFILTRATION]: The skill facilitates the capture and relay of sensitive NTLM authentication responses from victims to attacker-controlled targets, enabling unauthorized access to network resources and credentials.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to operate on data generated by network exploitation tools, creating a potential surface for indirect prompt injection.
Ingestion points: Network discovery and exploitation tool outputs (e.g., Responder logs, mitm6 responses, ntlmrelayx status messages) processed by the agent.
Boundary markers: No specific delimiters or "ignore embedded instructions" warnings are defined for the agent when processing tool outputs.
Capability inventory: The skill provides instructions for sophisticated network operations, credential manipulation, and remote command execution across multiple files (SKILL.md, COERCION_METHODS.md).
Sanitization: No validation or sanitization routines are provided for handling untrusted data received from the network during exploitation tasks.

ntlm-relay-coercion