Case study · Doktorsitesi.com
Jira × Slack Automation
A Claude Skill that turns Slack threads, Gmail, Loom recordings, and screenshots into structured Jira issues — with ADF bug reports, sprint routing, BE/FE splits, Airtable Bug List sync, and thread replies. Saves the product team roughly 10 hours per person every week.
Context & role
I designed and shipped this as Product & Project Specialist at Doktorsitesi.com: writing the jira-pm-skill-v3 spec, wiring Slack, Jira, and Airtable MCP flows, and rolling it out so PMs can route work to dev teams without manual ticket formatting.
The problem
PMs spent hours copying Slack screenshots into Jira by hand — inconsistent titles, missing test steps, wrong epics, and no trace in the Bug List. Bug reports lacked ADF structure, environment links, and BE/FE routing; Slack threads stayed open with no ticket link.
The solution
We built jira-pm-skill-v3: paste a Slack thread, Gmail, or Loom link and the skill classifies work type, picks epic/sprint, writes a full ADF description (error panels, test scenario, test accounts), creates the Jira issue, logs bugs to Airtable, and replies in the Slack thread with the issue link.
PM → Jira pipeline
Scroll into view to watch the automation: a PM input is parsed by the Claude Skill, routed to the right team and sprint, then lands as a Jira ticket with optional Airtable and Slack sync.
PM intake
- Slack threadBug report + screenshot
- GmailEmail chain
- LoomVideo recording
- PMPM confirmsSprint · assignee · priority
Skill pipeline
Claude Skill
jira-pm-skill-v3
Work type
BE / FE
ADF body
Sprint
Bug extras
Test scenario · env links · Airtable sync · Slack thread reply
Jira + sync
[FE] Uzman Panel - Takvim | Arama filtresi çalışmıyor
doktorsitesi.atlassian.net
- 1Skill parses intake & classifies work type
- 2Epic, team & ADF description built
- 3Jira issue created on DS project
- 4Airtable + Slack thread updated
What the skill handles
End-to-end PM routing — from messy chat intake to dev-ready Jira, without leaving Claude.
Intake
Reads Slack threads, Gmail chains, Loom transcripts, and panel screenshots — extracting repro steps, endpoints, and console/network data when available.
Routing
Classifies work type (Bug, Feature, Refinement…), assigns BE / FE / Mobil, resolves parent epic and live sprint, and splits BE+FE pairs with Blocks links.
Delivery
Creates ADF-formatted Jira issues, syncs Bug / Incident rows to Airtable Bug List, and posts a Slack thread reply tagging the reporter with the issue URL.
Outcomes
- ~10 hours saved weekly per PM on manual ticket writing
- Bug reports ship with test scenarios, env links, and test accounts
- Slack → Jira → Airtable stays linked for every bug intake
What's next
Want more context on how I automate PM workflows with Claude Skills, or to talk about a similar internal tooling problem?