The Agent Skills Directory

[PROMPT_INJECTION]: The skill documents patterns where variables from upstream projects are used directly in shell commands in downstream pipelines, creating a potential indirect injection surface. * Ingestion points: Variables such as $UPSTREAM_MR_REF and $K3S_CORE_MR_REF are passed from upstream triggers into downstream jobs in references/cross-pipeline-gating-mr.md and references/edge-infra-patterns-advanced.md. * Boundary markers: No specific boundary markers or validation steps are used when interpolating these variables into commands. * Capability inventory: The templates use these variables in high-privilege commands such as git clone, git fetch, and git checkout inside CI runner environments. * Sanitization: The provided examples do not include sanitization or validation of the upstream variables before execution.
[EXTERNAL_DOWNLOADS]: The skill references several external resources and tools required for CI/CD tasks. * Evidence: Fetches the release-cli tool from GitLab's official container registry (registry.gitlab.com/gitlab-org/release-cli:latest) and the yamllint tool from pipelinecomponents/yamllint:latest. * Evidence: Uses nix shell nixpkgs#yamllint to dynamically load the yamllint package from the Nixpkgs registry.

gitlab-ci