The Agent Skills Directory

[COMMAND_EXECUTION]: The skill contains instructions for a fallback validation path that involves executing a local script (validate_gitlab_ci.sh) to ensure the syntax and security of generated pipelines.- [PRIVILEGE_ESCALATION]: The instructions suggest using chmod +x to grant execution permissions to the local validation script, which is a standard operational step for tool scripts.- [EXTERNAL_DOWNLOADS]: The skill uses curl to interact with the well-known GitLab CI Lint API (https://gitlab.com/api/v4/projects/:id/ci/lint) for pipeline verification and supports retrieving official technical documentation to verify feature support.- [INDIRECT_PROMPT_INJECTION]: As a code generation tool that processes user requirements and external documentation, the skill possesses an injection surface where external content could influence agent output, although no exploitable pattern was detected.- [CREDENTIALS_UNSAFE]: The provided security guidelines explicitly warn against hardcoding secrets, providing educational examples of 'bad' practices (e.g., hardcoded tokens) and 'good' practices (e.g., using masked GitLab variables).

gitlab-ci-generator