The Agent Skills Directory

[COMMAND_EXECUTION]: The Kubernetes deployment template in SKILL.md includes the --insecure-skip-tls-verify=true flag. This disables SSL/TLS certificate validation for the kubectl command, which is a security risk that leaves the connection vulnerable to man-in-the-middle attacks.
[DYNAMIC_EXECUTION]: The 'Dynamic Child Pipelines' section in SKILL.md documents a pattern where a script (python generate_pipeline.py) generates a CI configuration file at runtime, which is then immediately executed via the trigger:include artifact directive.
[INDIRECT_PROMPT_INJECTION]: The dynamic pipeline pattern represents a surface for indirect instruction injection, as the agent context is influenced by dynamically generated content.
Ingestion points: The child-pipeline.yml artifact produced by the generate-pipeline job in SKILL.md.
Boundary markers: None; the template does not include delimiters or specific instructions to the pipeline runner to ignore untrusted content within the generated artifact.
Capability inventory: The pipeline uses the trigger capability to execute the instructions defined within the generated YAML file.
Sanitization: The skill does not demonstrate any validation, schema checking, or sanitization of the script's output before it is used to trigger a child pipeline.

gitlab-ci-patterns