Gitlab Cli Skills Publish

**Judgment:** High-quality, domain-focused reference skill for GitLab CLI (glab), well-suited for agents that operate via terminal and need to construct or reason about GitLab commands. It appears mature (3k+ downloads) and thoughtfully organized into sub-skills and workflows. **Strengths & Functionality** - Comprehensive coverage of glab domains: MRs, issues, CI/CD, repo, project management (milestones, iterations, labels, releases), auth/config (tokens, SSH/GPG keys), runners, clusters, OpenTofu, etc. - Includes practical workflows and decision trees (MR vs issue first, CI command selection, clone vs fork, project vs group labels) that help an agent choose the *right* command, not just any command. - Explicit awareness of glab v1.89.0 features, especially `--output json` / `-F json` options on many commands, which is valuable for structured parsing and automation. - Clear guidance on when to prefer glab versus the web UI, which can improve user experience and reduce misapplication of CLI operations. **Risks & Limitations** - The skill operates purely at the CLI level; actual success depends on the environment having `glab` installed and correctly authenticated, which is outside the skill’s control. - Some commands are inherently destructive or high-impact (e.g., merging MRs, deleting runners, modifying releases, managing tokens); an autonomous agent must apply additional safeguards, confirmation logic, or require explicit user consent. - The content is pegged to at least glab v1.89.0; over time, new glab versions may deprecate flags or add subcommands not covered here, so agents should not assume it is always perfectly in sync with the latest CLI. - Self-hosted GitLab instances can have configuration quirks (custom domains, disabled features, permissions differences) that may cause some suggested commands to fail or behave differently. **Recommended Scenarios** - Use when a user explicitly mentions GitLab CLI or `glab`, or when the agent is operating in a terminal environment and needs to perform GitLab actions programmatically (MRs, issues, CI, releases, runners, labels, etc.). - Prefer this skill for: - Constructing glab commands for daily dev workflows (issue → branch → MR → merge). - CI/CD diagnostics (pipeline status, failed jobs, logs, artifacts, retries). - Repository management actions (clone/fork decisions, repo view, auth setup). - Project-management activities via CLI (labels, milestones, iterations, releases). - Avoid relying on it as the *only* safety layer for destructive operations; pair with higher-level policies (confirmation prompts, dry-run strategies, or scope restrictions) when the agent is allowed to execute commands against real GitLab projects.