jira-digest

JIRA Digest

Use this skill when analyzing local JIRA transcript snapshots.

Scope

• Input is transcript text from slack-forge/transcripts/*jira*.md.
• Do not fetch JIRA/Slack data directly.

Event Types

• assignment
• status_change
• comment
• mention
• created
• priority_change
• sprint_change
• resolution

Extraction

For each event capture:

• ticket
• event_type
• summary
• needs_action

Prioritization

• Surface actionable items first.
• Group events by ticket.
• Summarize high-volume/noisy changes.

Actionability Rules

needs_action: true for:

• direct assignment to user/team
• explicit mentions/review requests
• blockers requiring intervention

Default informational events to false.

Confidence

• high: clear ticket reference with full event context
• medium: partial parse — ticket identified but event type or summary unclear
• low: ambiguous message, no clear ticket reference, or noise that could be JIRA-related

Digest Structure

Structure the digest content as an executive briefing:

1. Lead with actionable items — tickets needing user decisions, blocked items, assignments directed at the user. One paragraph per item explaining what and why.
2. Group informational events by category — status transitions (grouped by outcome: Done, In Progress, rework), assignments, comments. Use bullet lists.
3. End with "Key Tickets to Watch" — 3-5 strategically important tickets with a sentence of business context each (external dependencies, cross-team impact, product decisions).
4. Summarize high-volume noise rather than listing every event — "12 QA Subtasks created for sprint validation" is better than listing all 12.

Title format: "JIRA Digest — {date} ({timeframe})"

Never use "Untitled" as the title.

Provenance Requirements

For each digest record, preserve:

• source channel name and ID
• source author (bot username that posted the event)
• source timestamp (timestamp of first event in the digest window)
• transcript evidence mapping each jira_event entry to a specific message