The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: Fetches the Knative 'func' CLI binary and multiple Kubernetes deployment manifests from the official Knative organization repositories on GitHub.\n
Evidence: 'https://github.com/knative/func/releases/latest/download/func_linux_amd64' in SKILL.md.\n
Evidence: Multiple installation URLs targeting 'github.com/knative/' in scripts/install.sh.\n- [REMOTE_CODE_EXECUTION]: Downloads a binary ('func') from an external source and provides instructions to make it executable and run it, which is the standard installation method for this CLI tool.\n
Evidence: Download to 'func' then execute using 'chmod +x func' in SKILL.md.\n- [COMMAND_EXECUTION]: Utilizes 'kubectl' to deploy and manage cluster resources such as services, brokers, and triggers, which is the primary purpose of the skill.\n
Evidence: Extensive use of 'kubectl apply', 'kubectl patch', and diagnostic commands across all provided shell scripts.\n- [PRIVILEGE_ESCALATION]: Recommends the use of 'sudo' specifically for moving the downloaded 'func' binary into a system-wide executable path (/usr/local/bin).\n
Evidence: 'sudo mv func /usr/local/bin/' in SKILL.md.\n- [INDIRECT_PROMPT_INJECTION]: Several scripts interpolate user-provided arguments (like service name and image) into shell commands and YAML definitions without explicit sanitization, creating a surface for injection if provided with malicious input string.\n
Ingestion points: Arguments 'NAME' and 'IMAGE' in scripts/deploy-service.sh.\n
Boundary markers: None present.\n
Capability inventory: Cluster resource creation and modification via 'kubectl'.\n
Sanitization: Not implemented in the shell script.

knative