agentleFS
addxaiaddxai/enterprise-harness-engineering/skills/gitlab-mr/SKILL.md

gitlab-mr

Skill44 starsChanged 7 months agoApache-2.0View on GitHub

Create and manage GitLab Merge Requests on gitlab.example.com repositories, driving them to a truly mergeable state. Automatically generates User Story documentation, constructs MR descriptions (with blob links), pushes and creates MRs, polls CI pipeline status via API and fixes failures, resolves merge conflicts, and loops until all MR checks pass with no conflicts. Triggers when the user says "submit MR", "create MR", "push and create MR", "merge to main", or needs to submit the current branch for review.

  • Commits and pushes

4 other projects use a version of this.

Adopted by

Most recently taken up by

The file

Show its rules and text

We split this skill into the rules it gives an agent

14 rules. 0 are common, 11 are distinctive, and 3 are very specific to this project.

  • 0

    Common

    Most of the 3 edited versions we read keep these.

  • 11

    Distinctive

    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 an MR and drive it to a truly mergeable state (all CI green + no merge conflicts).

    line 8
  2. Distinctive:

    If there are uncommitted changes, commit or stash them first.

    line 38
  3. Distinctive:

    Check whether related documentation already exists in the repository:

    line 50
  4. Distinctive:

    If no related documentation exists, create one for this MR:

    line 57
  5. Very specific:

    Place the document at docs/plans/YYYY-MM-DD-<feature-name>.md (design doc) or docs/04-user-stories/<feature-name>.md (User Story)

    line 59
  6. Distinctive:

    After creating/updating the MR, wait for the pipeline to trigger and complete.

    line 123
  7. Distinctive:

    If the pipeline is still running, wait and check again:

    line 133
  8. Distinctive:

    If the pipeline fails, check which specific job failed:

    line 142
  9. Very specific:

    If the issue was in the MR description (check:mr-documentation): update the description then retry the job

    line 181
  10. Distinctive:

    If the issue was in code/docs: commit the fix, push, and a new pipeline will trigger automatically

    line 182
  11. Distinctive:

    After CI passes, also confirm the MR has no merge conflicts.

    line 194
  12. Distinctive:

    If there are merge conflicts, resolve them (rebase onto main), force push, and wait for the new CI to pass.

    line 201
  13. Very specific:

    If the pipeline doesn't trigger for a long time, you can manually run: glab ci run

    line 256
  14. Distinctive:

    After updating the MR description, you need to retry the failed job or wait for a new pipeline

    line 257

“Common” is measured: kept by at least 6 in 10 of the 3 edited versions of this file we could read. “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

GitLab MR

Create an MR and drive it to a truly mergeable state (all CI green + no merge conflicts). Core principle: write documentation first, then submit the MR, verify with APIs, and loop through fixes until the MR is mergeable.

Prerequisite: The current repository's remote points to gitlab.example.com, and glab CLI is installed.


Rules

  1. Never use the skip-doc-check label — every MR must have a real documentation link
  2. User Story must be relevant to the MR changes — CI checks relevance (check:mr-us-relevance); irrelevant links will be rejected
  3. Blob links must point to pushed files — the branch name and file path in the link must exist on the remote
  4. Verify with APIs, don't guess — confirm results at every step using glab commands

Execution Flow

Step 1: Check Prerequisites

# Confirm remote points to gitlab.example.com
git remote get-url origin

# Confirm current branch is not main/master
git branch --show-current

# Confirm no uncommitted changes
git status

If there are uncommitted changes, commit or stash them first.

Step 2: Ensure Documentation Exists

The MR description must contain documentation links. CI performs two checks:

  • check:mr-documentation: whether the link format is valid
  • check:mr-us-relevance: whether the linked document is relevant to the MR changes

Prefer GitLab blob links pointing to documentation files within the repository.

2a: Find or Create Documentation

Check whether related documentation already exists in the repository:

# Look for potentially related documents
ls docs/plans/ docs/04-user-stories/ 2>/dev/null

If no related documentation exists, create one for this MR:

  • Place the document at docs/plans/YYYY-MM-DD-<feature-name>.md (design doc) or docs/04-user-stories/<feature-name>.md (User Story)
  • The document content must cover the core changes in this MR (otherwise the relevance check will fail)
  • The document does not need to be long, but must describe: what was done and why

2b: Commit and Push Documentation

git add docs/plans/<doc-file>.md
git commit -m "docs: add <feature> design document"

Step 3: Push the Branch

git push -u origin <branch-name>

Step 4: Create or Update MR

Check if an MR already exists

glab mr list --source-branch <branch-name>

If no MR exists, create one

Analyze all commits in git log main..HEAD to write the MR title and description.

Blob link format: https://gitlab.example.com/<group>/<project>/-/blob/<branch>/<filepath>

  • <group>/<project> parsed from git remote get-url origin
  • <branch> is the current branch name
  • <filepath> is the in-repo path to the document
glab mr create \
  --title "<concise title, <70 chars>" \
  --description "$(cat <<'EOF'
## Change Summary
<Summary based on all commits, 2-3 bullet points>

## Related Documentation
- User Story (required): https://gitlab.example.com/<group>/<project>/-/blob/<branch>/docs/plans/<doc>.md
- Tech Design: <fill if available, otherwise remove this line>

## Change Type
- [x] <corresponding type>
EOF
)" \
  --push

If an MR exists but the description lacks documentation links

glab mr update <mr-id> --description "$(cat <<'EOF'
<complete description with blob links>
EOF
)"

Step 5: Wait and Check Pipeline

After creating/updating the MR, wait for the pipeline to trigger and complete.

# Wait for pipeline to start (usually takes a few seconds after push)
sleep 5

# Check pipeline status
glab ci status

If the pipeline is still running, wait and check again:

sleep 20
glab ci status

Step 6: Handle Failed Jobs

If the pipeline fails, check which specific job failed:

# Get status of all jobs in the pipeline
glab api "projects/:id/merge_requests/<mr-iid>/pipelines" | python3 -c "
import sys,json
pipelines = json.load(sys.stdin)
if pipelines:
    pid = pipelines[0]['id']
    print(f'Latest pipeline: {pid}, status: {pipelines[0][\"status\"]}')
"
# View failed job logs
glab api "projects/:id/pipelines/<pipeline-id>/jobs" | python3 -c "
import sys,json
for j in json.load(sys.stdin):
    if j['status'] == 'failed':
        print(f'FAILED: {j[\"name\"]} (id={j[\"id\"]})')
"
# Get the tail of a failed job's log
glab api "projects/:id/jobs/<job-id>/trace" 2>&1 | tail -30

Common Failures and Fixes

Failed JobCauseFix
check:mr-documentationMR description missing a valid documentation linkglab mr update <id> --description "..." to add blob links
check:mr-us-relevanceDocument content is not relevant to MR changesEdit the document to cover the MR's core changes, commit and push
validate:skillsSKILL.md format is non-compliantRun uv run python scripts/validate.py --skills to fix locally
validate:securitySecurity scan detected an issueRun uv run python scripts/validate.py --security to fix locally
Merge conflictBranch conflicts with maingit fetch origin main && git rebase origin/main, resolve conflicts and force push

After fixing:

  • If the issue was in the MR description (check:mr-documentation): update the description then retry the job
  • If the issue was in code/docs: commit the fix, push, and a new pipeline will trigger automatically
# Retry a failed job
glab ci retry <job-id>

# Or push a new commit to trigger a new pipeline
git push

Step 7: Confirm MR Is Mergeable

After CI passes, also confirm the MR has no merge conflicts. CI passing does not mean mergeable.

glab ci status        # Confirm CI is all green
glab mr view <branch> # Confirm no merge conflict

If there are merge conflicts, resolve them (rebase onto main), force push, and wait for the new CI to pass.

Loop through Steps 5-7 until both conditions are met: CI all green + no conflicts.

Once the MR is truly mergeable, report the MR URL to the user.


Examples

Full Workflow Example

# Create MR
glab mr create --title "feat: add retry logic" --description "..." --push

# Loop until CI passes
glab ci status  # -> if failed, check logs, fix -> check again

# After CI passes, check merge status
glab mr view feat/add-retry-logic
# -> conflicts? rebase to resolve -> force push -> wait for new CI

# Final confirmation: CI all green + no conflicts -> report MR URL

Bad/Good Examples

Bad — Declaring done just because CI passed

glab ci status  # -> success
# -> "MR !42 is created, CI passed!"  <- but there's actually a merge conflict, cannot merge

Bad — Posting an irrelevant link

glab mr create --description "User Story: .../README.md"  # unrelated to changes, CI will reject

Good — Driving to a truly mergeable state

# Create MR -> poll CI and fix failures -> check merge status -> resolve conflicts -> confirm mergeable

Notes

  • Branch names containing / (e.g., feat/my-branch) are correctly handled by GitLab in blob links
  • glab ci status shows the status of all jobs in the latest pipeline
  • If the pipeline doesn't trigger for a long time, you can manually run: glab ci run
  • After updating the MR description, you need to retry the failed job or wait for a new pipeline

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.