The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes system-level commands via PowerShell to manage security configurations.
Evidence: scripts/security_hardening_wrapper.ts utilizes child_process.spawn('powershell.exe', ...) to execute various administrative PowerShell scripts.
[PRIVILEGE_ESCALATION]: The skill is designed to modify system-wide registry keys and security policies, which necessitates administrative access.
Evidence: Scripts such as enforce_execution_policy.ps1 and setup_constrained_mode.ps1 perform modifications to the HKLM (HKEY_LOCAL_MACHINE) registry hive.
[INDIRECT_PROMPT_INJECTION]: The skill accepts user-supplied parameters that are subsequently interpolated into privileged command execution strings.
Ingestion points: The params object in security_hardening_wrapper.ts receives data from the agent context.
Boundary markers: No explicit boundaries or warnings are present in the scripts to isolate untrusted data.
Capability inventory: The skill can modify the system registry, alter execution policies, and perform code signing.
Sanitization: While PowerShell scripts utilize some parameter validation decorators, the TypeScript wrapper performs minimal sanitization before passing strings to the child process.
[EXTERNAL_DOWNLOADS]: The skill references established external services for timestamping during code signing operations.
Evidence: scripts/sign_module.ps1 and references/signing_guide.md interact with well-known certificate authorities such as DigiCert and Sectigo for timestamp verification.

powershell-security-hardening