atlassian-summarize

name	atlassian-summarize
description	Search Atlassian Confluence and Jira for content matching a keyword or statement, then produce a structured summary of all relevant findings

What I Do

Given a keyword, topic, or natural language statement from the user, I:

1. Search Confluence and Jira using atlassian_searchAtlassian with the provided query
2. Fetch full content for the top 3–5 most relevant results
3. Run follow-up searches with refined or related terms if initial results are sparse
4. Synthesize everything into a clean, structured summary

---

Workflow

Step 1 — Broad Search

Use atlassian_searchAtlassian with the user's keyword or statement verbatim as the query. Run multiple searches in parallel when the topic has multiple angles (e.g., "watchdog hardware" and "watchdog software").

Step 2 — Fetch Full Content

For Confluence pages: call atlassian_getConfluencePage with contentFormat: "markdown". For Jira issues: call atlassian_getJiraIssue or atlassian_fetchAtlassian with the ARI from search results. Fetch the top results in parallel to minimize latency.

Step 3 — Follow-Up Searches (if needed)

If initial results are too broad, too narrow, or reveal related subtopics, run 1–2 additional atlassian_searchAtlassian calls with refined terms derived from the first pass (e.g., a component name, subsystem name, or related feature discovered in the results).

Step 4 — Produce the Summary

Organize the output with the following structure:

## Summary: <keyword or statement>

### Key Findings
- Bullet-point distillation of the most important facts across all sources

### Sources Found
| Title | Type | URL |
|-------|------|-----|
| ...   | Page | ... |

### Detailed Breakdown
#### <Source Title>
- Core content summary (3–8 bullet points)
- Decisions, action items, or open questions recorded in this source

### Gaps / Open Questions
- Anything the sources did not answer or that appears unresolved

---

Rules

• Always use atlassian_searchAtlassian as the primary search tool unless the user explicitly says "use CQL" or "use JQL" — in that case use atlassian_searchConfluenceUsingCql or atlassian_searchJiraIssuesUsingJql instead.
• Fetch page/issue content in markdown format for cleaner output.
• Run independent fetches and searches in parallel to maximize efficiency.
• Prefer depth (full page content) over breadth (many shallow snippets).
• If the search returns both Confluence pages and Jira issues, cover both in the summary.
• Do not invent or guess URLs. Only link to URLs returned by the Atlassian tools.
• Keep the summary factual and objective — do not editorialize beyond what the sources say.

---

When to Use Me

Load this skill when the user asks to:

• "summarize everything about <topic> in Confluence"
• "collect all Atlassian info on <keyword>"
• "what do we know about <statement> across Jira and Confluence"
• "gather and summarize <topic> from Atlassian"