Jira CLI

Command reference and scripting patterns for jira-cli, the command-line tool for managing Jira issues, epics, and sprints without the web UI.

Prerequisites

Install with brew install ankitpokhrel/jira-cli/jira-cli, then run jira init once to authenticate.

These references assume jira on PATH is ankitpokhrel/jira-cli: a different tool can claim the same binary name. Before trusting any command below, run jira version once per session and confirm the output looks like this project (e.g. jira version 1.x.x with a Homepage: https://github.com/ankitpokhrel/jira-cli line). If it doesn't, stop and check which jira instead of guessing at flags.

For AI use, always add --plain (and usually --no-headers) to get parseable text output instead of the interactive table.

When to Use This Skill

Manual slash-command skill (/jira-cli): it does not auto-trigger on mentions of "jira" or "ticket". Invoke it directly for raw jira CLI commands (issues, epics, sprints) or scripting/automation. On tools that don't read Claude-Code frontmatter, disable-model-invocation is ignored: the /jira-cli phrasing itself is the manual-invocation signal there too.

Essential Commands

Curated top-7 for the most common operations. For anything else (filters, epic/sprint management, releases, output formats), load references/commands.md; don't try to recall the rest from memory.

# List recent issues (always use --plain for AI)
jira issue list --plain

# View issue details
jira issue view ISSUE-1 --plain

# Create an issue
jira issue create -tBug -s"Bug title" -yHigh -b"Description" --no-input

# Assign issue to yourself
jira issue assign ISSUE-1 $(jira me)

# Move issue to "In Progress"
jira issue move ISSUE-1 "In Progress"

# Add comment
jira issue comment add ISSUE-1 --comment "My comment"

# Add worklog
jira issue worklog add ISSUE-1 "2h" --comment "Implementation work"

Worked Examples

Confirm the CLI flavor before anything else:

$ jira version
jira version 1.5.1
Homepage: https://github.com/ankitpokhrel/jira-cli

A different jira version output (or a "command not found") means stop and resolve which jira before running any command from this skill.

Parse --plain output for scripting: columns are tab-separated, headers are on unless suppressed:

$ jira issue list -a$(jira me) -s"In Progress" --plain --no-headers --columns key,summary
PROJ-123	Fix login redirect loop
PROJ-124	Add rate limiting to API

jira issue list -a$(jira me) -s"In Progress" --plain --no-headers --columns key,summary | \
  while IFS=$'\t' read -r key summary; do echo "Working: $key - $summary"; done

Create non-interactively and capture the new key from the URL jira-cli prints:

$ jira issue create -tBug -s"Login redirect loop" -yHigh --no-input
Issue created
https://your-domain.atlassian.net/browse/PROJ-125

key=$(jira issue create -tBug -s"Login redirect loop" -yHigh --no-input | grep -oE '[A-Z]+-[0-9]+$')

Bulk sprint add: list the keys with a live filter, then add only the keys the user confirmed (see Destructive operations):

jira issue list -s"Ready for Dev" --plain --columns key,summary --no-headers
jira sprint add SPRINT_ID PROJ-123

Destructive operations

Applies to delete, bulk transition/assign/edit, removing issues from a sprint, and anything that touches more than one issue. Copy this checklist and tick it off:

- [ ] 1. List the exact keys with a read-only query
- [ ] 2. Show the keys and get explicit confirmation
- [ ] 3. Run the command once per confirmed key

1. Query read-only with the same filter the mutation would use, e.g. jira issue list -s"Ready for Dev" --plain --columns key,summary --no-headers.
2. Show that exact list to the user and wait for an explicit yes. If the list changes or the user edits it, return to step 1. No answer means stop and report.
3. Run the exact command for each confirmed key, e.g. jira issue delete PROJ-123. Never pipe the query into the mutation (... | xargs jira issue delete): nothing reviews what the pipe actually receives.

How to Use This Skill

Load reference files on demand, don't pull all three into context for a single command.

1. Comprehensive Commands Reference

Load: references/commands.md

Detailed syntax for issue management (list, create, edit, assign, move, view, link, clone, delete, comments, worklog), epic management, sprint management, releases, and output-format options.

2. Common Workflow Examples

Load: references/workflows.md

Multi-command patterns: sprint planning, code review handoff, bug triage, epic tracking, incident response, backlog grooming, cross-team coordination.

3. Scripting and Automation

Load: references/scripting.md

Raw bash automation that isn't already a curated skill: bulk assignment, auto-labeling, CSV export, velocity/metrics calculation, CI/CD hooks (GitHub Actions, GitLab CI, Jenkins), error handling and rate-limiting patterns.

Resources