status-report

name	status-report
description	Draft an audience-oriented Jira status report with atl. USE WHEN the user wants a weekly/daily progress narrative, blocker summary, or period report. DO NOT USE WHEN the output is a visual dashboard, direct Jira query, knowledge search, or codebase summary.

Status report with atl

Query Jira → analyze → shape for the audience → optionally publish. Confirm the scope before querying and always ask before publishing. Command details live in the jira and confluence skills.

Make export ATL_READ_ONLY=1 the first statement of every read-only Bash block so every later atl call and child process inherits the guard. Publishing is a separate explicitly approved block that removes the guard for one command.

Preflight: atl must be installed and configured. If command -v atl fails or a command exits 7 ("not configured"), run {{atl.setup_cmd}} and stop.

Workflow

1. Confirm scope

Ask only for missing information: scope kind (one/more epics, sprint/board, or whole project), key/id, time period (default: last 7 days), audience (team / manager / executive), and destination (chat only or Confluence).

2. Discover unfamiliar fields once

Do this only when the project uses unknown narrative, DoD, or risk fields. The inventory contains no values; choose promising exact names/ids before reading content.

export ATL_READ_ONLY=1
atl jira issue fields PROJ-1 --metadata-only

3. Choose the narrowest evidence path

For one or a few epics, prefer the aggregate evidence contract. Pass selected field names only when discovery found them:

export ATL_READ_ONLY=1
atl jira epic digest PROJ-1 --since 2026-07-01 --until 2026-07-07 \
  --status-field 'Delivery Notes' --dod-field 'Definition of Done'

Require top-level and named sources.*.complete; preserve every warning and staleness reason. For multiple epics, run one bounded digest per key.

For a sprint, resolve it and page the shared IssueList projection:

export ATL_READ_ONLY=1
atl jira sprint current --board <id>
atl jira sprint issues <sprint-id> \
  --columns position,key,summary,status,assignee,priority,updated --limit 50

Follow page.next_cursor until page.complete:true, or label the report partial. For a whole project, use explicit server-side buckets; do not fetch full issue bodies for counting:

export ATL_READ_ONLY=1
atl jira issue search --jql 'project = KEY AND statusCategory = Done AND resolved >= -7d' \
  --columns key,summary,status,assignee,priority,updated --limit 100
atl jira issue search --jql 'project = KEY AND statusCategory != Done' \
  --columns key,summary,status,assignee,priority,updated --limit 100
atl jira issue search --jql 'project = KEY AND priority in (Highest, High) AND statusCategory != Done ORDER BY priority DESC' \
  --columns key,summary,status,assignee,priority,updated --limit 100
atl jira issue search --jql 'project = KEY AND created >= -7d' \
  --columns key,summary,status,assignee,priority,created --limit 100

Status names vary per instance ("Blocked" is often a flag, not a status). Check returned values before building on them. For every IssueList inspect page.complete, page.truncated, and page.next_cursor; paginate or state the incomplete scope instead of treating a limit as absence.

4. Analyze

Compute done vs in-flight vs newly-created counts, notable completions, blockers with owners, stale high-priority risks, and unassigned work. Report what changed over the period, not raw issue lists. Separate observed facts from interpretation; never convert an incomplete source into zero or green status.

5. Shape for the audience

• Standup/team — terse Done / In progress / Blocked + owner / Next.
• Manager — summary, Highlights / Blockers & risks / Metrics / Next period.
• Executive — 3–5 sentences, overall RAG status, decisions needed; issue keys in a compact appendix rather than prose.

Every claim must trace to an issue key. Add an Evidence quality line naming incomplete/truncated sources and queries.

6. Publish — only after an explicit yes

env -u ATL_READ_ONLY atl conf page create --space KEY \
  --title 'Project X status — <date>' --parent <id> --from-md report.md

To refresh an existing page, pull it, edit the derived view, preview/apply the local merge, then use guarded push. Never publish merely because the user asked for a report; require a separate explicit yes for the destination and content.

7. Close the loop

Report the page id/title (or "not published"), JQL/digest periods used, source completeness, and any status/field names that did not exist and were approximated.