OpenParable
Skip to content
← Back to skills

Plan Issue

Productivity
plan-issue

Plan Issue is an Agent Skill for productivity published in the majiayu000/claude-skill-registry repository on GitHub: a SKILL.md instruction file an AI agent can load to do this task.

Explore the codebase and produce an implementation plan for a GitHub issue

View on GitHub ↗
QUICK START

How to use this skill

Bring this guide into your coding agent with a prompt tailored to the tool you use.

  1. Open your project in Codex.
  2. Copy and paste the prompt below.
  3. Review files and risks before approving any action.
Prompt to paste
Inspect the Agent Skill at https://github.com/majiayu000/claude-skill-registry/blob/HEAD/skills/planning/plan-issue/SKILL.md and all supporting files. Report provenance, permissions, scripts, dependencies, and security risks. Propose an exact installation into .agents/skills/plan-issue, but wait for my approval before writing files or running scripts.

Copying this prompt does not install or run the skill. Review third-party files before use.

Plan Implementation for GitHub Issue #$ARGUMENTS

Produce a detailed, approvable implementation plan before writing any code.


Step 1: Read the issue and announce work started

gh issue view $ARGUMENTS

Extract:

  • Goal / user-facing behaviour
  • Explicit requirements and acceptance criteria
  • Any referenced files, classes, or issues

Then post a comment to announce work has started:

gh issue comment $ARGUMENTS --body "🔧 **Work started** — exploring the codebase and drafting an implementation plan."

Step 2: Explore the codebase

Use Glob and Grep (or the Explore agent for broad searches) to locate:

  1. Files to create — identify the right package and naming convention by finding similar classes
  2. Files to modify — find every call site or injection point that needs updating
  3. Tests to create/update — find the existing test class for each file being modified
  4. Relevant patterns — read 1–2 existing similar classes to match structure (imports, annotations, error handling)

Key places to always check:

  • jhelm-core/src/main/java/org/alexmond/jhelm/core/ — action classes, models, interfaces
  • jhelm-kube/src/main/java/ — Kubernetes client implementation
  • jhelm-app/src/main/java/ — CLI commands (Picocli)
  • Matching test directories for each module

Step 3: Write the plan

Write the plan to /tmp/plan-$ARGUMENTS.md using this structure:

# Plan: <Issue Title> (Issue #$ARGUMENTS)

## Context
<2–3 sentences on what currently exists and what is missing>

## Files to Create
### 1. `path/to/NewClass.java`
- Purpose and key methods
- Important implementation notes

## Files to Modify
### N. `path/to/ExistingClass.java`
- What changes and why
- Code sketch if helpful

## Test Files to Create/Update
### N+1. `path/to/NewClassTest.java`
- Test cases to cover (success, error, edge cases)

## Verification
```bash
./mvnw test -pl <module>
./mvnw install
./mvnw spring-javaformat:apply && ./mvnw validate

Be specific: include exact class names, method signatures, and field types.
Flag any ambiguities or decisions that need the user's input before proceeding.

---

### Step 4: Post the plan as an issue comment

Post the plan contents to the issue so it's tracked on GitHub:
```bash
gh issue comment $ARGUMENTS --body "$(cat /tmp/plan-$ARGUMENTS.md)"

Step 5: Present and seek approval

Call ExitPlanMode to show the plan to the user for approval.

Do not write any production or test code until the plan is approved.

After approval, implement following the plan exactly, running /precommit between meaningful checkpoints.