# Create Backend Ralph Plans

Large Django features can lose task order, quality gates, and progress context. This skill creates a plan directory and Ralph prompt for iterative implementation.

## Install

```bash
npx skillstore add diversioteam/backend-ralph-plan
```

## Metadata

- Status: approved
- Slug: diversioteam-backend-ralph-plan
- Skillstore revision: r1
- Version status: missing
- Tree hash: dd3271b7a6153683417dcbb042d15f4c3ca9337259b8a57878620229e4ad4861
- Author: DiversioTeam
- GitHub username: DiversioTeam
- License: MIT
- Repository: https://github.com/DiversioTeam/agent-skills-marketplace/tree/main/plugins/plan-directory/skills/backend-ralph-plan
- Ref: c579efd02184f1128d6741d16e6c8d951d7d4208
- Supported tools: Claude, Codex, Claude Code
- Audit status: complete
- Agent install advisory: allowed
- Manual install advisory: allowed
- Artifact signature: available
- Audit attestation: unavailable
- Human verification: not\_verified
- Risk factors: external\_commands
- Quality score: 75
- Public page: https://skillstore.pages.dev/skills/diversioteam-backend-ralph-plan
- Manifest: https://skillstore.pages.dev/api/skills/diversioteam-backend-ralph-plan/manifest

## Capabilities

- Creates a PLAN.md task index with status, execution order, progress, and quality gate tables.
- Creates individual task files with goals, scope, dependencies, checklists, tests, completion criteria, and notes.
- Creates RALPH-PROMPT.md for repeated Ralph loop execution against the generated plan.
- Defines local verification commands for linting, type checks, Django checks, tests, coverage, and shortcut detection.
- Supports custom Django app paths, module paths, test filters, tooling commands, coverage targets, and iteration limits.
- Provides a complete user preferences example and reference templates for plan structure and quality gates.

## Use Cases

- Plan a Django Feature: Break a backend feature into ordered tasks with explicit verification gates and completion tracking.
- Prepare Autonomous Iteration: Create a Ralph prompt that can repeatedly resume work, check status, run tests, and commit completed tasks.
- Standardize Quality Gates: Capture consistent lint, type, test, coverage, Django, and shortcut checks for every task in a plan.

## Prompt Templates

### Create a Basic Plan

```
Create a backend Ralph plan for a Django feature named User Preferences. Use four tasks: model, service, API, and caching. Use app path accounts/ and test filter preferences.
```

### Add Tooling Details

```
Create a backend Ralph plan for Billing Webhooks. Use .bin/ruff check, .bin/ruff format, .bin/ty, .bin/pytest, and .bin/django check. Set coverage target to 90 percent.
```

### Include Dependencies

```
Create a backend Ralph plan for Audit Log Search. Include tasks for schema, indexing, service filters, API endpoints, permissions, and tests. Order tasks by dependency.
```

### Plan a Large Codebase Safely

```
Create a backend Ralph plan for Organization Settings in a large Django codebase. Add a warm-up orientation task before implementation and include blockers guidance.
```

## Limitations

- Designed for backend Django projects, not frontend projects or general application planning.
- Requires accurate user-provided task details, paths, test filters, and local tool commands.
- Does not implement the feature itself; it creates the plan files and execution prompt.
- Generated shell commands should be reviewed before running them in a repository.

## Best Practices

- Provide exact app paths, module paths, test filters, and local tooling commands before generating the plan.
- Keep each task focused on one backend layer or feature slice with four to eight checklist items.
- Review generated commands and task order before running the Ralph loop.

## Anti Patterns

- Do not use this skill for quick prototypes that do not need iterative verification.
- Do not create broad tasks that span many unrelated layers or require human decisions at every step.
- Do not run generated shell commands until their paths and arguments have been reviewed.

## Security Audit

- Audited at: 2026-07-05T13:10:15.915\+00:00
- Summary: The static findings are false positives from Markdown examples, inline code formatting, local developer command templates, and double-brace placeholders. I found no evidence of malware, credential access, unauthorized network activity, obfuscation, or prompt injection attempts in the reviewed files. The skill still expects users to review generated shell commands before running Ralph plans.

## Stats

- Views: 209
- Downloads: 12
- Favorites: 0
- Popularity score: 0