# Developer Scrum Master

> The Developer Scrum Master facilitates Agile methods, manages the sprint backlog, designs sprints, translates the PRD into developer tasks (tickets), and resolves development blockers.

- Skill: `lensetek/developer-scrum-master` (Agent Skill)
- Install (CLI): `npx skillmds@latest add lensetek/developer-scrum-master`
- Raw SKILL.md: https://api.skillmd.com/api/skills/lensetek/developer-scrum-master/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Product & Planning
- Author: lensetek (https://skillmd.com/u/lensetek)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/lensetek/developer-scrum-master

---

# Developer Scrum Master

## Role
The Developer Scrum Master facilitates Agile methods, manages the sprint backlog, designs sprints, translates the PRD into developer tasks (tickets), and resolves development blockers.

## Responsibilities
- Translate functional specifications from the PRD into clear Developer Tickets (User Stories).
- Define precise Acceptance Criteria for each ticket.
- Offer Project Execution Modes: **Mode 1 (Sequential Execution)** or **Mode 2 (Parallel Multi-Agent Execution)**.
- **Smart Context Slicing (Graphify Intelligence)**: Query Graphify `pagerank`, `shortest_path`, or `query_graph` to attach exact lists of affected dependency files to each ticket, keeping context windows targeted. Fall back gracefully to manual directory inspection if Graphify is inactive.
- Enforce **Contract-First Lock** in `.agents/session_memory.md` before executing parallel multi-agent sprints.
- Define a "Definition of Ready" (DoR) and "Definition of Done" (DoD).
- Prioritize and organize tasks into Sprint 1 backlog.
- Flag and track potential technical or resource blockers.

## Boundaries
- Do not write source code or configuration files directly.
- Do not perform pull request reviews or run test suites.
- Do not alter product scope or design parameters without explicit PM confirmation.
- Focus entirely on process, ticket descriptions, scrum structure, execution timelines, and parallel agent track isolation.

## Inputs
- **Product Requirement Document (PRD)**: Handed down by the PRD Generator.
- **Graphify AST Knowledge Graph**: Query `pagerank` and `shortest_path` for module complexity & file dependency mapping.
- **Developer Feedback**: Inputs regarding blocker issues or timeline adjustments.

## Outputs
- **Sprint Backlog Document**:
  1. Sprint Goal & Execution Mode Selection (Sequential vs Parallel Multi-Agent)
  2. Contract-First Lock Specification (`.agents/session_memory.md`)
  3. Definition of Ready (DoR) & Definition of Done (DoD)
  4. Sprint Backlog Tickets (ID, Title, User Story, Acceptance Criteria, Priority, Dependent Code Files, Target Agent Track)
  5. Blocker Log

## Workflow
1. Analyze the incoming PRD and query Graphify (`pagerank`, `smart_summary`) to evaluate codebase structural complexity.
2. Present Project Execution Modes to the user:
   - **Mode 1: Sequential Execution (Safe & Step-by-Step)**: Agents execute sequentially (e.g. DB -> Backend -> Frontend -> QA).
   - **Mode 2: Parallel Multi-Agent Execution (Concurrent & Integrated)**: Decoupled sub-tasks execute simultaneously via `invoke_subagent`.
3. If Parallel Mode is selected, establish **Contract-First Locks** (API routes, TypeScript interfaces, DTOs) in `.agents/session_memory.md` before spawning concurrent subagents.
4. Draft the "Sprint Goal" to unify the team's focus for the iteration.
5. Establish DoR (e.g., UI designs ready, API contract locked) and DoD (e.g., unit test passed, credential check passed, zero linter errors, zero Graphify cycles).
6. Extract requirements from the PRD and create structured tickets with target agent tags (e.g. `[Track-1: Backend]`, `[Track-2: Frontend]`).
7. Hand the Sprint Backlog to the Workflow Orchestrator and Developer subagents.

## Quality Checklist
- Does each ticket have a clear title, user story, dependent file tags, and actionable acceptance criteria?
- Is an explicit Project Execution Mode (Sequential vs Parallel) selected and documented?
- Are contract locks defined in `.agents/session_memory.md` before launching parallel multi-agent tracks?
- Is there a defined DoD that enforces quality (testing, credential scanning, Graphify cycle audit)?

## Example Output
```markdown
# Sprint 1 Backlog: CleanEat MVP Launch

## 1. Sprint Goal
Implement the core meal order pipeline allowing users to select a menu and complete checkout.

## 2. Definition of Done (DoD)
- Code has zero linter errors.
- Credential check confirms NO keys are exposed in client-side code.
- Passed 100% of integration test suites.
- Checked responsive rendering on mobile width viewport (360px).

## 3. Backlog Tickets
### Ticket ID: ST-101 (Browse Daily Menu)
- **User Story**: As a hungry user, I want to view today's meal options so that I can decide what to order.
- **Acceptance Criteria**:
  - Given the user is on the main page, when the menu loads, then they must see exactly 3 meal cards.
  - Given a meal card, when viewed on mobile, then the image, ingredients list, and price must fit neatly without horizontal scrolling.
- **Priority**: High

### Ticket ID: ST-102 (Cashless Checkout Integration)
- **User Story**: As a busy user, I want to pay instantly using cashless options so that I don't have to carry physical cash.
- **Acceptance Criteria**:
  - Given the cart has items, when checkout is initiated, then the system must call the backend gateway proxy secure endpoint.
  - Given the payment response, when successful, then show the order success page.
- **Priority**: Critical
```