home / skills / duc01226 / easyplatform / plan-ci

plan-ci skill

/.claude/skills/plan-ci

This skill analyzes Github Actions logs and delivers a collaborative implementation plan with multiple approaches and a recommended path.

npx playbooks add skill duc01226/easyplatform --skill plan-ci

Review the files below or copy the command above to add this skill to your agents.

Files (1)
SKILL.md
1.9 KB
---
name: plan-ci
description: "[Planning] Analyze Github Actions logs and provide a plan to fix the issues"
argument-hint: [github-actions-url]
infer: true
---

Activate `plan` skill.

> **CRITICAL:** Do NOT use `EnterPlanMode` tool — it blocks Write/Edit/Task tools needed for plan creation. Follow the workflow below.
> **Planning is collaborative:** Validate plan, ask user to confirm, surface decision questions with recommendations.

## Github Actions URL
 $ARGUMENTS

Use the `planner` subagent to read the github actions logs, analyze and find the root causes of the issues, then provide a detailed plan for implementing the fixes.

**Plan File Specification:**
- Every `plan.md` MUST start with YAML frontmatter:
  ```yaml
  ---
  title: "{Brief title}"
  description: "{One sentence for card preview}"
  status: pending
  priority: P1
  effort: {estimated fix time}
  branch: {current git branch}
  tags: [ci, bugfix]
  created: {YYYY-MM-DD}
  ---
  ```

**Output:**
Provide at least 2 implementation approaches with clear trade-offs, and explain the pros and cons of each approach, and provide a recommended approach.

## MANDATORY: Plan Collaboration Protocol (READ THIS)

- **Do NOT use `EnterPlanMode` tool** — it blocks Write/Edit/Task tools needed to create plan files and launch subagents
- **Do NOT start implementing** — plan only, wait for user approval
- **ALWAYS validate:** After plan creation, execute `/plan-review` to validate the plan
- **ALWAYS confirm:** Ask user to review and approve the plan using `AskUserQuestion` with a recommendation
- **ALWAYS surface decisions:** Use `AskUserQuestion` with recommended options for key architectural/design decisions
- **Planning = Collaboration:** The plan is shaped by user input — never treat it as a unilateral output
- Always plan and break many small todo tasks
- Always add a final review todo task to review the works done at the end
- Sacrifice grammar for concision. List unresolved questions at the end

Overview

This skill analyzes GitHub Actions logs and produces a collaborative, reviewable plan to fix CI failures. It reads the workflow run logs, identifies root causes, and drafts a plan.md with recommended fixes, trade-offs, and a sequence of small actionable tasks. The output is intended for review and approval before any implementation begins.

How this skill works

Provide a GitHub Actions run URL and the planner subagent will fetch and parse the logs, grouping failures by root cause and flakiness. It produces a plan.md that starts with the required YAML frontmatter, lists at least two implementation approaches with pros/cons, and breaks work into small todo tasks. The plan is then validated and presented for your feedback and approval; no changes are made until you confirm.

When to use it

  • CI runs are failing and logs contain unclear or multiple error types.
  • You need an actionable, reviewed plan before assigning fixes to engineers.
  • You want trade-offs evaluated for quick fixes versus long-term fixes.
  • You require small, testable tasks and a final review checklist for completed work.

Best practices

  • Supply the exact GitHub Actions run URL for accurate log parsing.
  • Keep plans focused: limit scope per plan.md to one failing workflow or related suites.
  • Prefer incremental fixes with clear rollback paths for high-priority pipelines.
  • List unanswered questions and decision points in the plan for collaborative resolution.
  • Estimate effort per task and include a final review/verification task.

Example use cases

  • A failing integration test suite where logs show database timeout and flaky network calls.
  • A new workflow failing due to missing secrets or changed environment variables.
  • Intermittent test failures suggesting flakiness that needs retries or test isolation.
  • Dependency updates causing build or restore errors across multiple jobs.

FAQ

What does the plan.md frontmatter look like?

The plan.md must start with YAML frontmatter including title, one-line description, status, priority, effort, branch, tags, and created date as specified.

Will this skill make changes automatically?

No. The skill creates a plan and validation step. Implementation only proceeds after you review and approve the plan.

How many approaches will you provide?

At least two distinct implementation approaches with clear pros and cons, plus a recommended approach.