jira-spec-doc

name	jira-spec-doc
description	Generates a product spec doc in Markdown from a Jira ticket. Use when the user provides a Jira ticket ID (e.g. CCC-284) or URL and asks for a spec, design doc, or tech spec ("write a spec for this ticket", "make a spec from this Jira").

Jira Spec Doc Generator

This skill takes a Jira ticket ID or URL and produces a comprehensive product spec doc as a Markdown file, following the Connect Spec Doc template with both a Design Doc and a Tech Spec section.

Step 1: Extract the ticket ID

Parse the ticket ID from whatever the user provides:

• A bare ID like CCC-284
• A full URL like https://dimagi.atlassian.net/browse/CCC-284
• A casual mention like "ticket 284 in CCC"

Step 2: Fetch ticket data from Jira

Use the Jira MCP tool (getJiraIssue) to fetch the ticket. Use cloudId dbff467f-3c3f-4ced-a2ba-a29e1941edd6 (Dimagi's Atlassian instance).

Pull these fields:

• summary — becomes the document title
• description — primary source of content
• issuetype.name — helps shape the narrative (Bug, Story, Task, Idea, etc.)
• status.name — for context
• assignee — for the Author field if present
• project.key and project.name — for the Epic/project field
• Any linked issues or parent epic if available

Step 3: Generate the spec doc

Write a Markdown file to ./outputs/<TICKET-ID>-spec-doc.md, creating the outputs/ directory if it doesn't exist.

Use the ticket data to populate the template as fully as possible. For fields the ticket doesn't supply, make a reasonable inference based on what you know, or leave a clear _[Add X]_ placeholder. The goal is a document that's genuinely useful as a starting point — not just a skeleton with empty fields.

Follow this exact structure:

---

Document structure

# <ticket summary>

**Jira Ticket:** [<TICKET-ID>](<ticket URL>)

---

## Design Doc

| Field | Details |
|---|---|
| **Related Materials** | [<TICKET-ID>](<ticket URL>) |
| **Mockups / Wireframes** | _Add link_ |
| **Author** | <assignee name, or "_Add name_"> |
| **Requested Reviewers** | _Add reviewers_ |
| **Feature Release Path** | Path 3 - Iterative Product Area |
| **Release Switch** | Required |
| **Implementation Tickets** | _Add tickets_ |

---

### Background

**Problem Description / User Story:**
<2–4 sentence description of the problem from the user's perspective, drawn from the ticket description>

**Background:**
<How did this problem arise? Why does it matter? What's been tried or observed? Infer from the ticket if not explicit.>

**Goals:**
- _Usage goals:_ <What measurable outcome are we aiming for after this ships?>
- _Implementation goals:_ <What must the implementation achieve to be considered done?>
- _(Optional) Effectiveness goals:_ <Are there effectiveness/workflow improvements we're targeting?>

**Non-Goals:**
- <What is explicitly out of scope for this phase?>

**Future Requirements:**
- <What related work might follow in a later phase?>

---

### Assumptions

- <Key assumption 1>
- <Key assumption 2>

---

### User Stories

**As a <role>:**
- <I want to X, so that Y.>

(Include one user story per affected role if multiple are apparent from the ticket.)

---

### Solution Summary

**Area 1 – <short label>**
<Description of the primary solution area>

**Area 2 – <short label, if applicable>**
<Description of a secondary solution area, if needed>

---

### Success Evaluation

<2–4 sentences describing how we'll know this was successful: metrics, behavioral changes, error rates, etc.>

---

### (Optional) Review Tracker

| Reviewer | Status | Notes |
|---|---|---|
| _Add reviewer_ | Not started | |

---
---

## Tech Spec

**Epic:** _Add Epic_
**Ticket Number & Description:** [<TICKET-ID>](<ticket URL>) — <ticket summary>

| Field | Details |
|---|---|
| **Related Materials** | [<TICKET-ID>](<ticket URL>) |
| **Mockups / Wireframes** | _Add link_ |
| **Author** | <assignee name, or "_Add name_"> |
| **Feature Release Path** | Path 3 - Iterative Product Area |
| **Release Switch** | Required |
| **Implementation Tickets** | _Add tickets_ |

---

### Introduction

<2–3 sentences summarizing what this ticket is about from a technical perspective, what the fix/change involves at a high level, and why it matters.>

---

### Solution

**Current Solution:**
<Describe the current behavior or system state that this ticket is addressing.>

**Proposed Solution:**
<Describe the proposed technical approach: what changes, where, and how. Include:
- Any external components or dependencies involved
- Security considerations
- How the solution scales
- Limitations of the approach
- Expected API changes
- Error logging considerations
- Any DB/model changes>

**Monitoring and Alerting Plan:**
<Describe any logging, analytics, or alerting that should accompany this change.>

**Deployment and Release:**
<Describe the release plan: feature flag usage, phased rollout if applicable, rollback strategy.>

**Alternative Solutions:**
<List 1–2 alternatives considered and why this approach was chosen over them.>

---

### Suggested Subheadings

**Model Changes:**
<Describe any data model or database schema changes, or state "No model changes anticipated.">

**Django View Changes:**
<Describe any backend view or API changes needed.>

**Frontend Changes:**
<Describe any frontend/UI changes, including component-level details where possible.>

**API Definitions:**
<Define any new or modified API endpoints: method, path, request/response format, permissions.>

---

### Further Considerations

- <Any open questions, edge cases, or cross-cutting concerns worth flagging for reviewers.>

---

Dimagi role and terminology glossary

Use these terms precisely and consistently throughout the spec doc. Abbreviations are commonly used in tickets — always expand them correctly:

Abbreviation	Full term
PM	Program Manager
NM	Network Manager
FLW	Front Line Worker

Never substitute generic terms like "user", "admin", or "project manager" when a specific role is implied. If a ticket says "PM-only action", write "Program Manager" throughout the doc.

Codebase references

Every Tech Spec must include a Relevant Codebases section immediately after the metadata table and before the Introduction. Link the correct repo(s) based on what the ticket involves:

• Web / backend work (Connect web app, PM portal, APIs, Django views, data models): commcare-connect
• Mobile work (FLW-facing Android app, OTP screens, PIN flows, mobile UX): commcare-android

Many tickets touch both — include both links when the work spans web and mobile. Use the ticket description to infer which codebase(s) apply: mobile keywords include "app", "Android", "OTP screen", "PIN", "FLW experience"; web keywords include "Connect web", "dashboard", "PM portal", "Django", "API endpoint".

Add this section to the Tech Spec right after the metadata table:

### Relevant Codebases

- [commcare-connect](https://github.com/dimagi/commcare-connect) — <one-line description of what changes here>
- [commcare-android](https://github.com/dimagi/commcare-android) — <one-line description of what changes here>

(Include only the repo(s) that are relevant; omit the other if it's truly not involved.)

Guidelines for writing good content

Draw from the ticket description thoroughly. Even a short description usually implies architecture, user roles, permissions, and failure modes. Reason through the implications rather than just paraphrasing.

Be specific about roles. Always use the full role names from the glossary above. Generic language ("users") is less useful than precise language ("Program Managers" or "Front Line Workers").

Make intelligent inferences. If the ticket says "hide button X for role Y," you can infer:

• There's currently no role check on button X (current state)
• A role check needs to be added (proposed solution)
• Backend enforcement may also be needed (security consideration)
• QA should test both roles (further considerations)

Leave clear placeholders. Anything you can't confidently infer should use _[Add X]_ or _Add X_ in italics so the author immediately knows what to fill in.

Keep structure consistent. The section headers must match the template exactly — reviewers expect to navigate to specific sections.

Output

Save the file to ./outputs/<TICKET-ID>-spec-doc.md, creating the outputs/ directory if it doesn't exist. Tell the user the file path where the spec doc was saved.