The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The script scripts/validate_prereqs.sh installs the Azure CLI using a shell-piped download command (curl | bash) targeting Microsoft's official infrastructure. This is a common pattern for installing required cloud dependencies from a well-known service.
[EXTERNAL_DOWNLOADS]: The skill downloads providers and modules from established registries (GitHub, HashiCorp, Sysdig) and fetches the Azure CLI installer from Microsoft. These operations target well-known technology providers and are consistent with the skill's primary purpose of infrastructure onboarding.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by ingesting data from locally persisted session files to pre-fill the discovery interview context.
Ingestion points: environment.yaml and customer-log.md are read at the start of each session in SKILL.md.
Boundary markers: No explicit delimiters or boundary instructions are used when interpolating this historical data into the current session context.
Capability inventory: The skill possesses high-privilege capabilities including terraform apply, helm install, and kubectl apply across multiple scripts.
Sanitization: No sanitization or validation logic is present for the data read from session history before it is used to influence the agent's wizard flow.

sysdig-onboarding