AgentSkillsCN

superplane-integration-issue-templates

在创建或审查 SuperPlane 集成问题内容时(包括基础集成问题,或组件触发器/动作问题)。当您需要生成问题模板、起草问题正文,或验证问题结构与规范时,此技能便能派上用场。

SKILL.md
--- frontmatter
name: superplane-integration-issue-templates
description: When creating or reviewing SuperPlane integration issue content (base integration issues or component trigger/action issues). Use when generating issue templates, drafting issue bodies, or validating issue structure and guidelines.

SuperPlane Integration Issue Templates

Use this skill when creating or reviewing issue content for SuperPlane integrations: base integration issues (one per tool) and component issues (triggers and actions, one per operation). Follow the integration-issue-conventions rule for IMPORTANT blocks, hierarchy, and title format.


Base integration issue

Purpose: One parent issue per tool. Establishes how the tool connects to SuperPlane. All component issues are children of this base.

Structure

  1. IMPORTANT block (at top) — use the base version: "Review and rethink the suggested connection method before implementing. If unsure, reach out on Discord first."
  2. Title: [{Integration Name}] Base (e.g. [GitHub] Base, [ArgoCD] Base)
  3. Description: 2–3 sentences on what the tool does and primary use cases. Always include Link: {URL}.
  4. Suggested Connection Method: Primary auth method (steps to generate credentials, required scopes, what to store in SuperPlane). Alternatives if applicable.
  5. Acceptance Criteria: has proper tests, documentation, code quality review, functionality review, ui/ux review.
  6. Follow up tasks: Once all components are done — announcement, outreach, marketplace/docs, templates/examples.
  7. Reference: Integration & component checklist.

Guidelines (base)

  • Connection method: Describe recommended auth (API Token, OAuth 2.0, App Installation, Service Account, Personal Access Token, Webhook Secret). Include where/how to generate credentials and required permissions/scopes.

Component issue (trigger or action)

Purpose: One issue per trigger or action. Each is a child of the base integration issue. Parent reference and IMPORTANT block at top (see integration-issue-conventions rule).

Structure

  1. IMPORTANT block (at top) — use the component version: "Review and rethink configuration options and output channels before implementing. If unsure, reach out on Discord first."
  2. Parent reference: **Parent (base integration):** #BASE_ISSUE_NUMBER (immediately after IMPORTANT block).
  3. Title: [{Integration Name}] {Operation Name} — Title Case; triggers use "On {Event}" (e.g. [GitHub] On Push, [ArgoCD] Sync Application).
  4. Priority: P1 - High / P2 - Medium / P3 - Low / P4 - Lowest (from agreed prioritization).
  5. Description: 1–2 sentences on what the component does to external systems, not that it exists. No "Enable X to Y as part of SuperPlane workflows."
  6. Use Cases: 2–3 specific, realistic scenarios (e.g. "Send Slack notification when production deployment pipeline completes"). Not generic "Automate {operation} in CI/CD workflow."
  7. Configuration: All fields with (required/optional), defaults, example values. Triggers = FILTERS only. Actions = EXECUTION PARAMETERS only.
  8. Outputs: Channel(s) with when emitted and what data. See Output channels below.
  9. Acceptance Criteria (optional): tests, documentation, reviews.
  10. Reference: Integration & component checklist.

Triggers vs Actions (critical)

  • Actions (execute operations): Configuration = execution parameters — what resource, how to perform, what data to send. Action config answers: "What should I DO?"
  • Triggers (listen for events): Configuration = filters — what events to listen for, what values to match (repository, status, labels). Trigger config answers: "What events should I LISTEN FOR?"

Wrong (trigger): [ArgoCD] On Sync Completed with config "Revision", "Prune" — those are for performing a sync, not filtering events.
Right (trigger): "Application Filter", "Sync Status" (Succeeded/Failed), "Health Status".

Output channels (components)

  • Triggers: Single default channel (outcomes determined downstream).
  • Actions: Depends on whether the user would model different workflow paths:
    • Multiple channels when the result should branch: e.g. success/failed, or clear/degraded/critical (e.g. by highest urgency), or approved/rejected.
    • Single default when there is one outcome stream or when "success" varies by user. Do not force success/failure on every action.
  • Channel names: Lowercase, single-word (success, failed, approved, timeout). Be consistent.

Anti-patterns to avoid

  • Generic use cases: "Automate {operation} within a CI/CD workflow", "Sync {service} changes into internal systems", "Create a workflow that reacts to {service} events."
  • Generic descriptions: "Enable {service} to {operation} as part of SuperPlane workflows.", "Trigger workflows in SuperPlane when {service} emits {event}."
  • Wrong config for triggers: Trigger with execution parameters (e.g. On Alarm with "Alarm Name", "Metric", "Threshold" for creating alarms). Use filter fields (Alarm Name Filter, State Transition, Severity).
  • Wrong title case: Create user → use Create User.
  • Vague example values: example, my-thing, test → use backend-api, prod-deployment, v1.2.3.

Validation checklist (before submitting)

  • Title uses Title Case.
  • Description explains WHAT happens to external systems.
  • Use cases are specific scenarios, not generic patterns.
  • Triggers: config has FILTER fields only.
  • Actions: output channels match whether user would branch (multiple vs default).
  • Configuration includes important API parameters; example values are realistic.