landscape-jira

name	landscape-jira
description	Creates well-formed Jira issues in the LNDENG project (warthogs.atlassian.net) — stories, epics, objectives, tasks, spikes, bugs, and more — with correct field mappings, sprint assignment, and acceptance criteria.

Cloud ID

Always pass warthogs.atlassian.net as cloudId to every Atlassian MCP tool call.

Issue Type IDs (LNDENG)

Type	ID	Hierarchy level	Notes
Theme	10386	3	Strategic, 1–3 year horizon
Objective	10390	2	OKR-level
Epic	10000	1	Must be broken into stories
Project-Risk	12391	1	Risk tracking
Project-Issue	12392	1	Materialized risks / blockers
Story	10002	0	Default work item
Task	10013	0	Small, distinct piece of work
Spike	10037	0	Research / investigation / prototyping
Bug	10015	0	Problem or error
Sub-task	10014	-1	Part of a parent task

Pass as: "issuetype": {"id": "<id>"}.

Typical hierarchy: Theme → Objective → Epic → Story/Task/Spike/Bug → Sub-task

Key Fields

Field	Field ID	Notes
Summary	summary	Required
Description	description	ADF format for API v3
Acceptance Criteria	customfield_10614	ADF format — NOT a plain string. Required for Epics/Objectives, recommended for Stories. See format below.
Story Points	customfield_10024	Number. Do NOT use customfield_10016 (legacy read-only)
Sprint	customfield_10020	Numeric sprint ID — plain integer, not an object. Not applicable to Epics/Objectives/Themes.
Assignee	assignee	{"accountId": "..."}
Parent	parent	{"key": "LNDENG-XXXX"} — any parent issue type
Issue Type	issuetype	{"id": "<id>"} — see table above
Labels	labels	Array of strings

Finding the Current Sprint ID

Never hardcode sprint IDs — they change every two weeks. Always look up the active sprint at call time:

project = LNDENG AND sprint in openSprints() ORDER BY created DESC

Read customfield_10020[0].id from any result. That integer is the sprint ID to pass.

For the next/upcoming sprint use sprint in futureSprints() instead.

Acceptance Criteria — ADF Format

customfield_10614 requires Atlassian Document Format (ADF), not a plain string. A plain string will error: "Operation value must be an Atlassian Document".

Bullet list template:

{
  "type": "doc",
  "version": 1,
  "content": [
    {
      "type": "bulletList",
      "content": [
        {
          "type": "listItem",
          "content": [{"type": "paragraph", "content": [{"type": "text", "text": "Criterion one"}]}]
        },
        {
          "type": "listItem",
          "content": [{      
        
      
    
  

Each bullet is a separate listItem. Do not use \n in text nodes — use multiple listItem or paragraph nodes instead. Literal \n characters render as the string \n in Jira, not a line break.

Sprint Field Format

Sprint must be a plain integer, not an object:

"customfield_10020": 30952

{"id": 30952} will error: "Number value expected as the Sprint id."

Description Field

Plain strings work via the MCP wrapper. For multi-paragraph content, use ADF — multiple paragraph nodes, never \n literals inside text nodes.

Useful JQL Patterns

Find open stories assigned to you in the current sprint:

project = LNDENG AND sprint in openSprints() AND assignee = currentUser() AND resolution = Unresolved

Find child stories under a parent:

project = LNDENG AND parent = LNDENG-XXXX ORDER BY created ASC

Find stories created by you recently:

project = LNDENG AND reporter = currentUser() AND created >= -7d ORDER BY created DESC

Writing Good Acceptance Criteria

AC is about building the right product. DoD is about building it right. Don't conflate them.

AC are required for Epics and Objectives, recommended for Stories.

Three types of AC

• User-centered — does it meet user needs? Functionality, experience, edge cases. Written from the user's perspective.
• Operational — does it meet internal standards? Security, scalability, compliance, maintainability. Unlike DoD (task-level quality), these are deployment-readiness criteria tied to operational use cases.
• System-centric — does it work in the broader ecosystem? APIs, shared services, downstream dependencies, indirect users.

Qualities of good AC

• Testable — can you verify it against a proposed implementation or in QA?
• Result-oriented — focus on outcomes for the user, not the implementation. "Latency < 250ms" not "written in Go".
• Unambiguous — understood the same way by PM, engineering, and design.
• Concise — enough detail to make the right implementation choices, not a full spec.

What AC is not

• Not a description of the UI or solution — describe what the user can do, not how it's built.
• Not a DoD checklist (unit tests, docs, CI passing) — those belong in the PR template or DoD.
• Not a exhaustive list — if you're writing many ACs for one story, break the story up.

Format tips

• Use plain declarative statements: "User can X", "System must Y", "Returns Z on error"
• Group by type when there are many: user-centered first, then operational, then system
• State inputs, outputs, states, error conditions, and edge cases where relevant
• Leave implementation choices open — describe the outcome, not the mechanism

Gotchas

1. AC is ADF, not a string — wrap in doc > bulletList > listItem > paragraph > text. Most common failure.
2. Sprint is a plain integer — not {"id": ...}.
3. Story Points — customfield_10024. The field customfield_10016 is read-only legacy, don't write to it.
4. Parent field — use parent: {"key": "LNDENG-XXXX"}. Works for epics and objectives. customfield_10014 is the read-only epic link.
5. No \n in text nodes — use separate ADF nodes for line breaks; \n renders as a literal backslash-n.
6. AC vs description — AC belongs in customfield_10614 only. Description = context/approach. Never duplicate AC into the description.
7. New issues land in "Untriaged" — transition them explicitly if a different status is needed.