
from claude-skills28
Create or update a Drupal merge request (MR), prepare required AI-disclosure text at policy threshold, and surface verify/review status and RTBC guidance before
Creates or updates a Drupal merge request and prepares the AI disclosure required by project policy. The skill automates the routine MR workflow: confirms branch and target, checks verification and review preconditions, assesses whether AI authored a significant portion of the change, crafts disclosure text for the issue and MR description, and reports MR status and RTBC guidance. It wraps existing CLI tooling (drupalorg-cli) rather than reimplementing network flows and emphasises safe gates and validation.
Use this skill when a contributor asks to submit or update a Drupal patch or MR (e.g. via /drupal-ai-contrib:submit). It should run after code changes and verification steps are available. If verify or review hasn’t been run or is stale, the skill surfaces that fact and lets the contributor decide whether to proceed. It is not a moral arbiter — it enforces policy workflow and disclosure placement, not acceptance.
Designed for server-side assistant tooling that can call CLIs and agent task tools (Claude-derived agents or similar automation runners). The skill expects access to shell tools like drupalorg and knowledge-layer lookups (dev-guides-navigator).
A Drupal contribution submission worker skill that guides creating/updating merge requests and handling AI-disclosure requirements. Well-structured SKILL.md with detailed preconditions, procedure, and troubleshooting. References external scripts (reverify-list.sh, review-mark.sh) and delegates to other skills (drupal-gitlab) but ships none itself. No scripts to run. Validates identifiers against explicit patterns and warns about write-token scope — solid security awareness. Niche audience: Drupal contributors using AI-assisted workflows.
No bundled scripts to test. Security is good — explicit input validation patterns for identifiers, warnings about write-token scope, no credential storage. The skill delegates MR creation to another skill rather than doing it directly, which is architecturally clean but means it can't operate independently. Usefulness is limited to Drupal contributors specifically working with AI-assisted contributions — a niche but real audience.