agentleFS
pawellachpawellach/pm-ai-prompts/skills/jira-intake/SKILL.md

jira-intake

Skill0 starsChanged 2 months agoMITView on GitHub

Analytical ticket intake before handing off to a developer. Use when — (1) user writes `/jira-intake PROJ-123` (single ticket), (2) user writes `/jira-intake PROJ-1 PROJ-2 PROJ-3` (up to 5 tickets at once), (3) user says "analyse this ticket", "do intake for PROJ-123", "assign ticket to dev", (4) user writes `/jira-intake stats` (registry summary). Gathers context from Jira and Confluence, writes an analytical comment with implementation scope, transitions to Waiting, and reassigns to chosen person. Batch mode: analyses in parallel, shows a combined summary, asks for assignment once for all. Do NOT use for tickets where a dev is already working (status In Progress).

2 other projects use a version of this.

Adopted by

Most recently taken up by

The file

Show its rules and text

We split this skill into the rules it gives an agent

26 rules. About 0 are common, 15 are distinctive, and 11 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈15

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 11

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Call getAccessibleAtlassianResources.

    line 17
  2. Distinctive:

    Stop — do not continue without MCP.

    line 29
  3. Distinctive:

    Ask in one message:

    line 41
  4. Very specific:

    Call getTransitionsForJiraIssue(cloudId, sampleTicket) → match status names to transition IDs.

    line 51
  5. Very specific:

    Before every intake run (except stats):

    line 64
  6. Very specific:

    If invoked as /jira-intake stats:

    line 79
  7. Very specific:

    If missing or empty → "No intake history yet.

    line 82
  8. Distinctive:

    Stop after showing report — do not continue to intake.

    line 109
  9. Distinctive:

    If history file found in step 0c → prepend to analysis:

    line 132
  10. Very specific:

    Check $registry.tickets for other open tickets (status != "Done") that share:

    line 140
  11. Distinctive:

    If any found → add to comment (step 3):

    line 145
  12. Distinctive:

    If a local codebase is available: grep for class/object names mentioned in the ticket.

    line 160
  13. Distinctive:

    If not available — skip silently.

    line 161
  14. Distinctive:

    Use any available MCP tools (database schema lookups, API docs, etc.) if relevant to the ticket domain.

    line 165
  15. Distinctive:

    If complexity = Complex, append at the end of the comment:

    line 198
  16. Very specific:

    If config.transitions.waiting is set: transitionJiraIssue(cloudId, ticket, transition: { id: config.transitions.waiting })

    line 209
  17. Distinctive:

    If null — skip this step.

    line 212
  18. Very specific:

    Before showing assignment options, check $workload:

    line 220
  19. Distinctive:

    Ask: "Who should this be assigned to?

    line 226
  20. Distinctive:

    After all analyses (steps 1–4) show a table:

    line 234
  21. Very specific:

    After response: for each ticket → lookupJiraAccountId → editJiraIssue → transitionJiraIssue(inProgress) — all in parallel.

    line 252
  22. Distinctive:

    After assignment is confirmed:

    line 258
  23. Very specific:

    Read current ~/.claude/skills/jira-intake/registry.json (or start fresh).

    line 262
  24. Distinctive:

    Add/update entry for each processed ticket:

    line 263
  25. Very specific:

    Write back to ~/.claude/skills/jira-intake/registry.json.

    line 277
  26. Distinctive:

    If the file already exists, append a new dated section (don't overwrite):

    line 293

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

jira-intake — Analytical ticket intake (v2)

Role: Analyst / PO. Prepares ticket(s) for handoff to a developer.


Step 0 — Check MCP, config, and state

0a. Check Atlassian MCP

Call getAccessibleAtlassianResources. If unavailable:

Atlassian MCP plugin is not active. To use jira-intake:

1. Open Claude Code → Settings (⚙) → Plugins
2. Enable: "Atlassian" (official plugin)
3. Sign in with your Atlassian account
4. Restart Claude Code
5. Run /jira-intake again

Stop — do not continue without MCP.

0b. Check config

Check for ~/.claude/skills/jira-intake/config.json.

  • Exists and complete → use it.
  • Missing or incomplete → run Setup below.

Setup (first use or /jira-intake setup)

  1. getAccessibleAtlassianResources → if one instance, use its url as cloudId; if multiple, ask which to use.
  2. Ask in one message:
    I need to know your Jira status names and a sample ticket number.
    
    What is the status name when:
    1. A developer starts working? (e.g. "In Progress", "In Development")
    2. A ticket is waiting to be assigned / queued? (e.g. "Waiting", "Ready for Dev", "Backlog")
       Type "none" if you don't have this status.
    3. Please also provide a sample ticket number (e.g. PROJ-123)
  3. Call getTransitionsForJiraIssue(cloudId, sampleTicket) → match status names to transition IDs.
  4. Save to ~/.claude/skills/jira-intake/config.json:
    {
      "cloudId": "<url>",
      "statuses": { "inProgress": "<name>", "waiting": "<name or null>" },
      "transitions": { "inProgress": "<id>", "waiting": "<id or null>" }
    }
  5. Confirm: "Config saved. You can now use /jira-intake PROJ-123."

0c. Load state (registry + history)

Before every intake run (except stats):

  1. Registry — read ~/.claude/skills/jira-intake/registry.json if it exists. Parse into memory as $registry. If missing, treat as { "version": 1, "tickets": {} }.
  2. History check — for each ticket being processed, check if ~/.claude/skills/jira-intake/history/<TICKET>.md exists. If yes, read it → attach as extra context in step 2 with header:
    [Re-run] Previous analysis from <date>: <summary from history file>
  3. Workload snapshot — from registry, compute for each developer: count of tickets where status != "Done" and lastRun within last 7 days. Store as $workload[name] = N.

Stats mode (/jira-intake stats)

If invoked as /jira-intake stats:

  1. Read ~/.claude/skills/jira-intake/registry.json.
  2. If missing or empty → "No intake history yet. Run /jira-intake PROJ-123 to start."
  3. Otherwise, generate report:
📊 jira-intake registry stats

Total tickets processed: N
Last run: YYYY-MM-DD

Complexity breakdown:
  Simple:  N (XX%)
  Medium:  N (XX%)
  Complex: N (XX%)

Top assignees:
  1. Name — N tickets
  2. Name — N tickets
  3. Name — N tickets

Top epics:
  1. EPIC-XX — N tickets
  2. EPIC-XX — N tickets
  3. EPIC-XX — N tickets

Active (not Done, last 7 days): N tickets

Stop after showing report — do not continue to intake.


Single vs. batch mode

  • Single ticket (/jira-intake PROJ-123) → run full sequence (steps 1–6) for that ticket.
  • Multiple tickets (/jira-intake PROJ-1 PROJ-2 ...) → run steps 1–4 in parallel for each ticket, then step 5 combined (one summary table + one assignment question for all).

Max 5 tickets at once. If more provided, inform and take the first 5.


Step 1 — Read ticket

getJiraIssue(fields: ["summary", "description", "comment", "reporter", "assignee", "issuetype", "priority", "status", "labels", "components"]) with responseContentFormat: "markdown".


Step 2 — Gather context (parallel per ticket)

2a. History context

If history file found in step 0c → prepend to analysis:

[Re-run: poprzednia analiza z <date>]
Assignee: <name> | Complexity: <X> | Summary: <1-line>

2b. Cross-ticket intelligence

Check $registry.tickets for other open tickets (status != "Done") that share:

  • Same epic as the current ticket
  • Same assignee (if already assigned)
  • Same component keywords (match 2+ words from title/description)

If any found → add to comment (step 3):

**Kontekst: powiązane otwarte**
- PROJ-X: <title> (assignee: <name>, complexity: <X>)

2c. Confluence

Extract keywords from ticket description. searchConfluenceUsingCql: text ~ "keywords" — top 3 results.

For relevant results: getConfluencePage → extract relevant sections.

2d. Local repository (optional)

If a local codebase is available: grep for class/object names mentioned in the ticket. If not available — skip silently.

2e. Other tools (optional)

Use any available MCP tools (database schema lookups, API docs, etc.) if relevant to the ticket domain.


Step 3 — Write analytical comment in Jira

addCommentToJiraIssue — no confirmation needed.

**Pre-analysis**

**Business context**
[1-3 sentences — what the user wants to achieve]

**Affected areas**
- [Area 1]: [relevant files/classes/modules found]
- [Area 2]: [if applicable]
- Confluence: [links to relevant pages]

**Implementation scope**
[Concrete action items for the developer]

**Complexity estimate:** Simple / Medium / Complex
[1 sentence justification]

**Open questions / risks**
[Only if something requires a decision before dev starts]

[If cross-ticket context found:]
**Kontekst: powiązane otwarte**
- PROJ-X: <title> (assignee: <name>)

If complexity = Complex, append at the end of the comment:

💡 Ten ticket ma złożoność Complex. Chcesz pełną analizę wymagań?
Uruchom: /requirements-analyst <TICKET>

Step 4 — Transition to Waiting

If config.transitions.waiting is set: transitionJiraIssue(cloudId, ticket, transition: { id: config.transitions.waiting })

If null — skip this step.


Step 5 — Assignment

Workload warning

Before showing assignment options, check $workload:

  • For each candidate assignee with $workload[name] > 3 → annotate in the table: ⚠️ <name>: N otwartych

Single mode

Ask: "Who should this be assigned to? (provide name or email)"

Show workload warning inline if applicable.

Take the answer → lookupJiraAccountId by email or name → editJiraIssue (assignee) → transitionJiraIssue(config.transitions.inProgress).

Batch mode — combined summary

After all analyses (steps 1–4) show a table:

Analysis complete.

| Ticket   | Title (short)           | Complexity | Notes                        |
|----------|-------------------------|------------|------------------------------|
| PROJ-1   | ...                     | Simple     | UI change                    |
| PROJ-2   | ...                     | Complex    | Multiple service layers      |
| PROJ-3   | ...                     | Medium     | API integration              |

Workload (last 7 days, open tickets):
  Damian: 2 | Bartek: 1 | Marcin: 4 ⚠️

Who should each ticket be assigned to?
(e.g. "PROJ-1 → alice@company.com, PROJ-2 → bob, PROJ-3 → alice")

After response: for each ticket → lookupJiraAccountId → editJiraIssue → transitionJiraIssue(inProgress) — all in parallel.


Step 6 — Update state (registry + history)

After assignment is confirmed:

6a. Update registry

Read current ~/.claude/skills/jira-intake/registry.json (or start fresh). Add/update entry for each processed ticket:

{
  "SAL-XXXX": {
    "lastRun": "YYYY-MM-DD",
    "assignee": "<name>",
    "complexity": "Medium",
    "epic": "<epic key or null>",
    "status": "In Progress"
  }
}

Write back to ~/.claude/skills/jira-intake/registry.json.

6b. Write history file

Write ~/.claude/skills/jira-intake/history/<TICKET>.md:

# <TICKET>
**Date:** YYYY-MM-DD
**Assignee:** <name>
**Complexity:** Simple / Medium / Complex
**Summary:** <1-sentence summary of what was analysed>
**Analysis:**
<full text of the Jira comment written in step 3>

If the file already exists, append a new dated section (don't overwrite):

---

## Re-run: YYYY-MM-DD
**Assignee:** <name>
**Complexity:** <X>
**Summary:** <summary>
**Analysis:**
<full comment text>

Step 7 — Confirm

✅ Intake complete

PROJ-1 → Alice Smith (In Progress)
PROJ-2 → Bob Jones (In Progress)
PROJ-3 → Alice Smith (In Progress)

Registry updated. History saved for: PROJ-1, PROJ-2, PROJ-3

Error handling

  • Confluence / repo unavailable → skip that section, continue.
  • lookupJiraAccountId returns no results → ask for exact email.
  • Transition fails → log error, continue to reassignment.
  • Config corrupted → re-run setup.
  • Registry write fails → log warning, continue (don't block intake).
  • History dir missing → create ~/.claude/skills/jira-intake/history/ before writing.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.