Jira Administration
SkillProductivityRun bounded Jira administration workflows for project Components or Atlassian instance migration. Use for jira-components, sync Jira components, jira-instance-migration, changed Jira URL, or repoint Jira. Both modes are sealed behind read-first analysis and explicit user approval before any Jira, credential-session, or repository mutation.
Available today. Use it from your connected AI after setup.
No other account needed.
Connect ahel once, and every AI you use reads what you have installed.
Then ask your AI: use the Jira Administration skill
What this skill tells your AI
The instructions your AI receives, as published by upex-galaxy/agentic-qa-boilerplate in .agents/skills/jira-administration/SKILL.md and read by ahel’s review.
Mode routing
Choose exactly one mode and load only its reference.
| Mode | Legacy alias / trigger | Reference |
|---|---|---|
components | jira-components, reconcile/sync Jira Components | references/components.md |
instance-migration | jira-instance-migration, changed/repoint Jira instance | references/instance-migration.md |
If the mode is unclear, ask. Never combine both modes in one run.
Compact Rules
The sealed mutation contract. Binding on every run of either mode:
- Exactly ONE mode per run:
components(references/components.md) orinstance-migration(references/instance-migration.md). Load only that mode's reference. Never combine the two, never fall through into the other. - Mode unclear → ASK. Do not infer one from a bare "fix Jira" / "sync Jira" request.
- Load
/aclibefore any Jira operation. Load other tool-owner skills only when the selected reference requires them. - Missing MCP or Jira credentials = HARD STOP (
AGENTS.mdCritical Rule #10). Name the exact env var, point at.env/.env.example, ask for an agent-session restart. No workaround, no partial run. - Read-first on every mutation: inspect the live state before authoring any plan. Nothing is created, applied, deleted, or repointed without the user's explicit approval given inside the same run.
components: derive and inspect → author the plan file → dry-run → WAIT for explicit approval → only then--apply.instance-migration: resolve and confirm BOTH instances → audit and verify reachability → WAIT for explicit approval → only then change files or theaclisession. That session lives at~/.config/acliand is machine-global: re-login repoints every repo on the host, not just this one.- The Atlassian host lives in
.agents/project.yaml→issue_tracker.atlassian_urland NOWHERE else locally. A staleATLASSIAN_URLin.envor the process environment is contamination to DELETE, never to update — a second copy is what goes stale. - Template-repo carve-out: if
.agents/project.yaml→project.project_nameisnull, the repo is an un-onboarded template. Leaveatlassian_urlandproject_keynull, say so in the report, and never manufacture a commit to hide the emptiness. - Run only the selected reference's verification steps. Never run the other mode's.
- Forward
$ARGUMENTSunchanged.
Read full SKILL.md when: the mode is ambiguous, a dry-run diff or migration audit looks wrong, or you need the selected reference's step-by-step phases and verification list.
Signals
- GitHub stars
- 21
- Forks
- 12
- Last commit
- Sep 2026
ahel review
S4info
community integration — published by upex-galaxy, not jira
Automated review, not a security audit. Ruleset v1.
Advanced
- Catalog kind
- skill
- Gateway key
jira-administration- Source
- github.com/upex-galaxy/agentic-qa-boilerplate