The Agent Skills Directory

[SAFE]: The skill provides documentation and workflow instructions for the legitimate GitLab command-line interface (glab). It encourages context-aware command usage and proper authentication practices.
[SAFE]: External references and documentation links point exclusively to official GitLab documentation (docs.gitlab.com) and well-known repositories (e.g., salsa.debian.org, gitlab-org/cli).
[INDIRECT_PROMPT_INJECTION]: The skill interacts with external data provided by GitLab instances (such as issue descriptions, merge request content, and API responses) and provides the agent with capabilities to execute commands based on this data.
Ingestion points: Data enters the agent's context through commands like glab mr view, glab issue view, glab repo view, and glab api endpoints.
Boundary markers: The skill includes instructions to confirm targets before performing write operations if the repository or host context is ambiguous, which serves as a manual guardrail.
Capability inventory: The agent is granted capabilities to perform shell-based operations including repository management, CI/CD pipeline control, and merge request updates using the glab CLI.
Sanitization: The instructions do not specify custom sanitization for external content, relying instead on the default behavior of the glab tool.

glab-cli