The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill includes functionality to ingest and analyze Windows Event Logs, which represents an indirect prompt injection surface.
Ingestion points: The analyze_event_log method in scripts/agent.py accepts and processes event data provided as a list of dictionaries.
Boundary markers: Absent. There are no delimiters or instructions to the agent to ignore embedded commands within the log data.
Capability inventory: The skill can write files to the local system (save_deployment in agent.py) and perform sensitive Active Directory modifications via generated PowerShell scripts.
Sanitization: Absent. The event data is processed without explicit sanitization or validation of the content against malicious instructions.
[COMMAND_EXECUTION]: The skill generates and executes PowerShell scripts that perform high-privilege Active Directory operations.
Evidence: The New-HoneytokenAdmin, Add-HoneytokenSPN, and New-DeceptiveBloodHoundPath functions in scripts/Deploy-ADHoneytokens.ps1 create users, register SPNs, and modify DACLs/SACLs on sensitive objects.
[DYNAMIC_EXECUTION]: The Python agent dynamically generates executable PowerShell scripts from internal templates.
Evidence: The PowerShellGenerator class in scripts/agent.py assembles script content using string interpolation and writes it to disk for subsequent execution.

deploying-active-directory-honeytokens