| name | ralph |
| description | Convert existing PRDs to Linear projects and issues for the Ralph autonomous agent system. Use when you have an existing PRD markdown file and need to set up Ralph. Triggers on: convert this prd, turn this into ralph format, set up ralph for this prd, ralph convert. |
Ralph PRD to Linear Converter
Converts existing PRD markdown files to Linear projects and issues for autonomous execution by Ralph.
Prerequisites: Linear MCP must be configured. See https://linear.app/docs/mcp
The Job
- Read the PRD markdown file provided by the user
- Parse user stories and requirements
- Select a Linear team (interactively)
- Create a Linear project with PRD content
- Create Linear issues for each user story
- Save
.ralph-project configuration
Step 1: Parse the PRD
Read the PRD file and extract:
- Title/Name from
# PRD: [Name] or first heading
- Description/Overview from Introduction section
- Goals from Goals section
- User Stories from User Stories section (parse US-XXX format)
- Non-Goals from Non-Goals section
- Technical Considerations from that section
Step 2: Select Linear Team
Use mcp__linear-server__list_teams to get available teams and ask the user which team to use.
Step 3: Create Linear Project
Use mcp__linear-server__create_project:
{
"name": "<PRD Title>",
"team": "<team-id-or-name>",
"description": "Branch: ralph/<feature-kebab-case>\n\n<Full PRD content as markdown>",
"state": "planned"
}
Critical: The first line of the description MUST be Branch: ralph/<feature-name> for Ralph to identify the correct git branch.
Step 4: Create Issues from User Stories
For each ### US-XXX: section in the PRD:
- Extract title (text after
US-XXX:)
- Extract description (the "As a..." line)
- Extract acceptance criteria (the
- [ ] items)
- Map position to priority (1st=Urgent, 2nd=High, etc.)
Create issue using mcp__linear-server__create_issue:
{
"title": "US-001: <Title>",
"team": "<team-id-or-name>",
"project": "<project-id-or-name>",
"description": "<As a... description>\n\n## Acceptance Criteria\n<Criteria as checklist>",
"priority": 1,
"state": "Todo"
}
Priority Mapping
| PRD Story Position | Linear Priority |
|---|
| 1 (first) | 1 (Urgent) |
| 2 | 2 (High) |
| 3 | 3 (Normal) |
| 4+ | 4 (Low) |
Step 5: Save Configuration
Create .ralph-project in the Ralph directory:
{
"linearProjectId": "<project-id>",
"branchName": "ralph/<feature-name>"
}
Story Size: The Number One Rule
Each story must be completable in ONE Ralph iteration (one context window).
Ralph spawns a fresh Claude Code instance per iteration with no memory of previous work. If a story is too big, the LLM runs out of context before finishing and produces broken code.
Right-sized stories:
- Add a database column and migration
- Add a UI component to an existing page
- Update a server action with new logic
- Add a filter dropdown to a list
Too big (split these):
- "Build the entire dashboard" - Split into: schema, queries, UI components, filters
- "Add authentication" - Split into: schema, middleware, login UI, session handling
- "Refactor the API" - Split into one story per endpoint or pattern
Rule of thumb: If you cannot describe the change in 2-3 sentences, it is too big.
Story Ordering: Dependencies First
Stories execute in priority order. Earlier stories must not depend on later ones.
Correct order:
- Schema/database changes (migrations)
- Server actions / backend logic
- UI components that use the backend
- Dashboard/summary views that aggregate data
Wrong order:
- UI component (depends on schema that does not exist yet)
- Schema change
Acceptance Criteria: Must Be Verifiable
Each criterion must be something Ralph can CHECK, not something vague.
Good criteria (verifiable):
- "Add
status column to tasks table with default 'pending'"
- "Filter dropdown has options: All, Active, Completed"
- "Clicking delete shows confirmation dialog"
- "Typecheck passes"
- "Tests pass"
Bad criteria (vague):
- "Works correctly"
- "User can do X easily"
- "Good UX"
- "Handles edge cases"
Always include as final criterion:
"Typecheck passes"
For stories with testable logic, also include:
"Tests pass"
For stories with testable business logic (optional):
"Tests written first (TDD)"
Use this when the story involves complex logic, algorithms, or API contracts that benefit from test-first development.
For stories that change UI, also include:
"Verify in browser"
(The agent will auto-select Playwright MCP or dev-browser based on availability)
Validation Before Creating
Before creating in Linear, verify:
If stories are too big, split them before creating issues.
Output Summary
After creating everything, show the user:
Converted PRD to Linear:
Project: <Project Name>
ID: <uuid>
Branch: ralph/<feature-name>
Issues Created:
- <TEAM>-123: US-001: <Title> (Urgent)
- <TEAM>-124: US-002: <Title> (High)
- <TEAM>-125: US-003: <Title> (Normal)
- <TEAM>-126: US-004: <Title> (Low)
Saved .ralph-project configuration.
To start Ralph:
./ralph.sh [max_iterations]
Example
Input: tasks/prd-task-status.md
Creates:
-
Linear Project: "Task Status Feature"
- Description contains full PRD with
Branch: ralph/task-status on first line
-
Linear Issues:
TEAM-45: US-001: Add status field to tasks table (Urgent)TEAM-46: US-002: Display status badge on task cards (High)TEAM-47: US-003: Add status toggle to task list rows (Normal)TEAM-48: US-004: Filter tasks by status (Low)
-
Local file: .ralph-project with project ID and branch name
Splitting Large PRDs
If a PRD has big features, split them:
Original:
"Add user notification system"
Split into:
- US-001: Add notifications table to database
- US-002: Create notification service for sending notifications
- US-003: Add notification bell icon to header
- US-004: Create notification dropdown panel
- US-005: Add mark-as-read functionality
- US-006: Add notification preferences page
Each is one focused change that can be completed and verified independently.
Checklist
Before finishing:
