Jira Cloud Skill Pack Files

Jira

Issue tracking, sprint planning, and project management on Atlassian Cloud. Agents drive the <workspace>.atlassian.net web UI. Jira Server / Data Center self-hosted instances diverge enough in DOM that selectors here are not guaranteed to work there.

Browser Access

Requires the user to be signed in to Jira/Atlassian in the connected browser. SSO sessions persist across runs. Set JIRA_BROWSER_URL in environment variables to the workspace base URL (https://acme.atlassian.net) so workflows do not have to hardcode it.

Tool	Usage
navigate	Go to board, issue, search, or filter URL
snapshot	Read board state, issue panel, search results
click	Buttons, links, issue cards, transition options
type	Forms, search fields, JQL input

Project types

Jira projects come in two flavors that share the issue model but expose different UI:

Type	URL	Notes
Company-managed (classic)	/jira/projects/<KEY>	Shared schemes, JQL, custom workflows
Team-managed (next-gen)	/jira/software/projects/<KEY>	Per-project config, simpler workflow

Both publish the same /browse/<KEY>-<num> URL for individual issues, so issue navigation works the same regardless of project type.

Boards, backlog, and sprint

A Software project exposes:

Surface	URL suffix	Purpose
Board	/boards/<id>	Active sprint or kanban columns
Backlog	/backlog	Refinement queue, sprint planning
Reports	/reports	Burndown, velocity
Issues	/issues	Project-scoped issue search
Pages	/pages	Linked Confluence pages

Drag-and-drop on the board moves cards between columns and triggers the workflow transition associated with the destination column. Drag actions on the backlog reorder by rank.

Issues

An issue at /browse/<KEY>-<num> exposes:

• Summary (title, editable inline)
• Description (Atlassian Document Format — rendered, editable inline)
• Status pill (click to transition)
• Sidebar: Assignee, Reporter, Priority, Labels, Components, Fix versions, Story points, Sprint, Linked issues, Attachments, Web links
• Activity feed: Comments, History, Work log

Issue keys are stable across renames and project moves — they are the canonical identifier.

JQL

JQL (Jira Query Language) powers all advanced search. Filter examples:

JQL	Result
project = SCRUM AND status = "In Progress"	All in-progress in SCRUM
assignee = currentUser() AND resolution = Unresolved	My open work
sprint in openSprints() AND project = SCRUM	Active sprint contents
created >= -7d AND project = SCRUM	New this week
labels = "needs-review"	Tagged for review
parent = SCRUM-123	Subtasks of an epic/story
text ~ "search phrase"	Full-text search

JQL is entered at /issues/?jql=… (URL-encode the query). Saved filters live at /issues/?filter=<id>.

Transitions

Workflow transitions are surfaced in two places:

1. The status pill at the top of an issue — clicking opens the available transitions for the current status.
2. The board column drop target — releasing a dragged card runs the column's transition.

Available transitions depend on the workflow scheme attached to the project; they are not universal. Always read the current status and available transitions from the issue page before scripting a move.

Linking

Issues link via:

• Hierarchy: Epic ↔ Story / Task ↔ Subtask. Epic link sits in the sidebar.
• Issue links: typed (blocks, is blocked by, relates to, duplicates, is duplicated by, causes, is caused by). Visible in the Linked Issues section.
• Web links: free-form URL + title. Useful for Slack permalinks, GitHub PR URLs, Figma frames.

Components, labels, versions

Three orthogonal taxonomies:

• Components — pre-defined, project-scoped, often map to teams or services.
• Labels — free-form, global, lowercase-hyphenated by convention (needs-review, tech-debt, customer-impact).
• Fix versions — release identifiers, project-scoped.

Use components when an admin should curate the taxonomy. Use labels when anyone should add new tags ad hoc.

Conventions

• Always include: summary, description, acceptance criteria.
• Past tense in bug descriptions ("Button failed to load").
• Link to source context (Slack thread, page URL, log line).
• Assign to the right team — don't leave unassigned.
• Don't reopen resolved tickets — create a new one and link with relates to / caused by.

Priority

Level	When to use
P1 / Highest	Production outage, data loss
P2 / High	Broken feature affecting users
P3 / Medium	Enhancement or improvement
P4 / Low	Cosmetic or nice-to-have

Common patterns

View issue: navigate to /browse/<KEY>-<num>, snapshot for full state.

Search: navigate to /issues/?jql=<encoded>, snapshot results table.

Transition: click the status pill, choose target transition, fill required fields if the workflow demands them.

Create issue: use the jira_browser_create_issue workflow or click Create in the top bar — fill summary, project, type, then submit.

Gotchas

• The issue page is heavily SPA — fields auto-save on blur; scripting must wait for the save toast before moving on.
• Sidebar fields like Sprint and Epic only appear on Software projects; Service / Business projects expose different fields.
• JQL is case-insensitive on field names but case-sensitive on string values.
• The board view caches sprint contents — after a transition, refresh before reading the new column state.
• /browse/<KEY> (without number) redirects to the project; double-check that the issue key is fully qualified.
• API-style endpoints (/rest/api/3/…) are not part of this browser pack — use the Jira REST API directly when JSON output matters.