teamcraft-jcg:stakeholder-update

---
name: teamcraft-jcg:stakeholder-update
description: Generate an audience-appropriate stakeholder update by pulling live Jira sprint data, GitHub PR/pipeline status, and Confluence artifacts. Produces human-facing communication from AI-native project artifacts — status reports, client updates, executive summaries, engineering updates. Works in Claude Cowork and Claude Code.
argument-hint: "[audience type and update period — e.g. 'client weekly update' or 'executive sprint summary']"
disable-model-invocation: true
user-invocable: true
allowed-tools:
  - Bash
  - mcp__sooperset-mcp-atlassian__jira_get_all_projects
  - mcp__sooperset-mcp-atlassian__jira_search
  - mcp__sooperset-mcp-atlassian__jira_get_agile_boards
  - mcp__sooperset-mcp-atlassian__jira_get_sprints_from_board
  - mcp__sooperset-mcp-atlassian__jira_get_sprint_issues
  - mcp__sooperset-mcp-atlassian__confluence_search
  - mcp__sooperset-mcp-atlassian__confluence_get_page
  - mcp__sooperset-mcp-atlassian__confluence_get_page_children
  - mcp__sooperset-mcp-atlassian__confluence_create_page
---

## Goal

Produce a stakeholder update that gives the audience exactly what they need — no more. Different audiences need fundamentally different information: a client cares about delivered value and risk to their timeline; an executive cares about status and decisions needed; an engineering team needs specifics about what shipped, what's blocked, and what changed. The source of truth is always the live project data — not memory, not estimates.

## Hard Constraints

- Pull live data from Jira before composing any update. Never estimate or describe sprint status from memory.
- The update is generated FROM the project artifacts, never instead of them. If the PRD or tech decisions are stale, flag it — do not paper over it in the update.

## Identify the Project and Sprint

Use `mcp__sooperset-mcp-atlassian__jira_get_all_projects` to surface available projects and ask the user which one this update covers. Then use `mcp__sooperset-mcp-atlassian__jira_get_agile_boards` to find the project's board and `mcp__sooperset-mcp-atlassian__jira_get_sprints_from_board` to identify the active or most recently completed sprint. Confirm with the user.

## Pull Live Data

From Jira, gather the current state of the sprint:
- Use `mcp__sooperset-mcp-atlassian__jira_get_sprint_issues` for the sprint's complete issue list with status, assignee, and issue type
- Supplement with `mcp__sooperset-mcp-atlassian__jira_search` using JQL for additional detail: priority, labels, created/updated dates
- Identify issues that are at risk: blocked, overdue, or showing no recent activity
- Sprint goal from the sprint data

From GitHub (via Bash, if available):
- PRs: `gh pr list --state all --json number,title,state,createdAt,mergedAt,headRefName` filtered to the sprint period
- Pipeline status: `gh run list --limit 20` for recent workflow run outcomes
- If Bash is unavailable, omit GitHub data and note the gap

From Confluence: ask if the user can point at the PRD and any other relevant documents for context. Load them if available with `mcp__sooperset-mcp-atlassian__confluence_get_page`.

## Determine Audience and Compose

Use `$ARGUMENTS` and context to determine the audience type. Ask if not clear. The audience shapes every word:

**Client / external stakeholder** — delivered value in business terms, risks to scope or timeline with mitigations, decisions the client needs to make. No internal jargon. No ticket numbers. Green / Yellow / Red overall status with plain-language rationale.

**Executive / leadership** — overall health (G/Y/R), progress toward sprint goal, risks that need a decision or escalation, asks. Under 300 words. No ticket-level detail.

**Engineering team** — what shipped, what's blocked and why, what changed in scope or priority, what's up next. Ticket-level specifics welcome. Technical language appropriate.

**Cross-functional partners** — what affects them: scope changes, new dependencies on their team, timeline shifts, decisions they need to weigh in on.

See `references/example-stakeholder-update.md` for structure and tone calibration by audience type.

## Review and Store

Show the complete update for review. Offer to adjust tone, length, or level of detail.

Ask whether to store this in Confluence. If yes, ask which space. Use `mcp__sooperset-mcp-atlassian__confluence_search` to surface available spaces if needed. Once confirmed, use `mcp__sooperset-mcp-atlassian__confluence_get_page_children` to browse structure and recommend a parent page. Create with `mcp__sooperset-mcp-atlassian__confluence_create_page`.

## Done

Share the update text and Confluence page URL if stored. The update is derived from the live project state — rerun this skill any time the audience or period changes.