The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the subprocess.run function to interact with the oc command-line tool. This is used to verify the user's session status and to retrieve transient authentication tokens via oc whoami -t. The implementation follows security best practices by passing arguments as a list and avoiding the use of a shell for execution.
[EXTERNAL_DOWNLOADS]: The script fetches job metadata and container image tags from prow.ci.openshift.org and quay.io. These operations are performed to assist the user in selecting the correct job parameters and are limited to official Red Hat/OpenShift infrastructure.
[DATA_EXPOSURE]: Accesses the OpenShift CI configuration file at ~/.config/openshift-ci/kubeconfig to authenticate API requests. The scripts are designed to handle credentials securely in memory, ensuring that tokens are not leaked into terminal output or logs.
[INDIRECT_PROMPT_INJECTION]: The skill processes data from external web sources, creating a potential surface for indirect injection.
Ingestion points: Fetches and parses data from prow.ci.openshift.org and quay.io.
Boundary markers: The workflow structures this information into markdown tables for the agent's context.
Capability inventory: The skill can perform authenticated POST requests to trigger CI jobs and execute oc commands.
Sanitization: The script uses specific regular expressions to validate job names and version tags before they are displayed or used.

rhdh-prow-trigger