agentleFS
vladcheckvladcheck/agentic/skills/gitlab-eas-tickets-publish/SKILL.md

gitlab-eas-tickets-publish

Skill1 starsChanged 10 days agoMITView on GitHub

Use when publishing an approved EAS spec and dependency-ordered tickets to GitLab, optionally with an MR. Not for standalone specs or tickets.

1 other project uses a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

8 rules. About 0 are common, 5 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈5

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Create or update GitLab objects only within the scope the user requested.

    line 21
  2. Distinctive:

    Stop here when no implementation MR was requested.

    line 28
  3. Very specific:

    Create a dedicated branch from fresh origin/main in a separate worktree, leaving any occupied or dirty primary clone untouched:

    line 32
  4. Distinctive:

    Run the project's local checks before pushing.

    line 40
  5. Distinctive:

    Create the MR without auto-merge:

    line 42
  6. Very specific:

    Use glab issue list --output json --jq ..., not --json.

    line 52
  7. Very specific:

    Use --description-file <file> for issue and MR bodies to preserve Markdown and avoid shell-quoting errors.

    line 53
  8. Distinctive:

    Before retrying a write after an interrupted or unclear response, query GitLab for the current state and reuse any object already created.

    line 56

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Publish an EAS specification, tickets, and implementation MR

Coordinate the complete GitLab publishing workflow while keeping the detailed spec and ticket procedures in their dedicated skills.

Select the workflow

  • For a spec-only request, use publish-spec-gitlab-milestone and stop.
  • For a ticket-only request, use publish-tickets-gitlab and stop.
  • For the combined workflow, use publish-spec-gitlab-milestone first, then publish-tickets-gitlab, and continue here only if the user requested an implementation MR.

Prerequisites

  • Confirm that the user approved the exact specification and dependency-ordered ticket breakdown. Never publish an unapproved draft or quiz version.
  • Resolve <repo>, <project-id>, the milestone title, and the triage label from the target project. Use <issue-iid> and <mr-iid> for project-scoped IIDs. Use <milestone-id> only for the global numeric REST milestone id, never for an IID.
  • Match the language of the approved artifacts and project conventions. Keep code identifiers and branch names in the repository's established language.
  • Create or update GitLab objects only within the scope the user requested.

Publish and verify

  1. Publish the approved specification through publish-spec-gitlab-milestone. Retain the milestone title, <milestone-id>, and anchor <issue-iid> for the remaining steps.
  2. Publish the approved breakdown through publish-tickets-gitlab. That workflow creates issues strictly in dependency order so every blocker has a real <issue-iid> before its dependents are created.
  3. Retain the ticket-to-issue mapping, verification evidence, and unblocked frontier returned by the ticket workflow. Do not rerun glab issue create or infer IIDs from ticket numbers. Treat ## Blocked by references such as - #<issue-iid> (<title>) as the portable dependency contract recorded by that workflow.
  4. Stop here when no implementation MR was requested. Otherwise, continue with the approved implementation scope.

Open an implementation MR

  1. Create a dedicated branch from fresh origin/main in a separate worktree, leaving any occupied or dirty primary clone untouched:

    git fetch origin main
    git worktree add -b <branch> <worktree> origin/main
  2. Work through git -C <worktree> ... rather than changing the caller's directory. Implement only the approved issue scope.
  3. Run the project's local checks before pushing. Record the exact commands and their results. Resolve failures before pushing; never claim unrun or failed checks as passing.
  4. Push <branch> and write the MR description to a file. Include a concise implementation summary, Closes #<issue-iid>, and the exact verification commands and results.
  5. Create the MR without auto-merge:

    glab mr create --repo <repo> --source-branch <branch> --target-branch main --title "<mr-title>" --description-file <file> --yes
  6. Return the <mr-iid> and verification summary. Leave the MR open for review. Never enable auto-merge. Never merge the MR in this workflow; a merge requires separate, explicit user approval.

GitLab CLI safeguards

  • Use glab issue list --output json --jq ..., not --json.
  • Use --description-file <file> for issue and MR bodies to preserve Markdown and avoid shell-quoting errors. Use --yes for noninteractive creation.
  • glab api has no --jq flag in the supported CLI contract; pipe API JSON to jq when filtering is needed.
  • Address milestone REST reads and updates as projects/<project-id>/milestones/<milestone-id> with the global numeric <milestone-id>.
  • Before retrying a write after an interrupted or unclear response, query GitLab for the current state and reuse any object already created. Treat remote state as the source of truth and avoid duplicate issues, milestones, or MRs.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.