feat: add incident-commander plugin with /create-postmortem command

[Matovidlo]

Summary

Adds a new Incident Commander plugin with a /create-postmortem slash command that automates incident post-mortem creation. The command gathers incident information from a Slack channel and creates a structured post-mortem document in Confluence based on a template.

Key changes:

• Created new plugins/incident-commander/ plugin structure (separate from developer plugin)
• Added comprehensive /create-postmortem command with detailed section-by-section instructions
• Added MCP Server Setup section to main README with Atlassian and Slack configuration
• Plugin includes Atlassian MCP pre-configured in plugin.json

The command acts as a senior engineer helping an incident commander and:

• Reads complete Slack channel history including all threads
• Fetches the post-mortem template from Confluence (template pageID: 3568304146)
• Creates/finds year directories under parent page (pageID: 3568009242)
• Generates structured post-mortem with: Overview, Impact, Timeline, Action Items, Responders, etc.
• Uses Europe/Prague timezone by default
• Asks for clarification when information is missing

Prerequisites for users:

• Atlassian MCP configured (for Confluence)
• Slack MCP configured (for reading channel messages)

Updates since last revision

• Created new incident-commander plugin - Moved /create-postmortem out of developer plugin into its own dedicated plugin
• Expanded command instructions with detailed guidance for each post-mortem section:
  • Overview, What Can We Learn, What Happened, Impact, Responders, Timeline
  • What Went Well, What Didn't Go So Well, Action Items, Messaging, Runbooks
• Updated Confluence page IDs: Parent page 3568009242, Template 3568304146
• Changed timezone from UTC to Europe/Prague (configurable)
• Added plugin README with configuration table and best practices

Review & Testing Checklist for Human

• Verify Confluence page IDs - Parent page 3568009242 and template 3568304146 must exist and be accessible
• Verify Atlassian MCP URL - Confirm https://mcp.atlassian.com/v1/sse is the correct endpoint
• Test Slack token extraction steps - Follow the browser instructions to extract XOXD/XOXC tokens
• Verify Slack MCP tool names - The command assumes tools like mcp__slack__list_channels exist
• Verify post-mortem template structure - The detailed section instructions should match your actual Confluence template headings

Recommended test plan:

1. Install the incident-commander plugin: /plugin install incident-commander
2. Run claude mcp add --transport sse atlassian https://mcp.atlassian.com/v1/sse
3. Run /mcp to authenticate with Atlassian
4. Extract Slack tokens from browser following the README instructions
5. Run the Slack MCP setup command with your tokens:

   claude mcp add --transport stdio slack-mcp-server \
     --env SLACK_MCP_XOXC_TOKEN=<your-xoxc-token> \
     --env SLACK_MCP_XOXD_TOKEN=<your-xoxd-token> \
     -- npx -y slack-mcp-server

6. Run /mcp to verify Slack MCP is working
7. Create a test incident channel with some messages
8. Run /create-postmortem <test-channel-name>
9. Verify the post-mortem is created in Confluence with correct structure

Notes

• The Confluence page IDs are organization-specific defaults that users should update for their own setup
• The Slack token extraction instructions are browser-specific (Chrome/Firefox) and may need updates if Slack changes their web app
• The post-mortem section instructions are detailed but may need adjustment to match your actual Confluence template structure
• Generated post-mortems are a starting point - human review is essential before finalizing
• Timezone defaults to Europe/Prague; update in the command for other organizations

Link to Devin run: https://app.devin.ai/sessions/d04a77c2722c4fa8af64b0476ceed2b5
Requested by: Martin Vasko (@Matovidlo)

[devin-ai-integration]

🤖 Devin AI Engineer

I'll be helping with this pull request! Here's what you should know:

✅ I will automatically:

• Address comments on this PR. Add '(aside)' to your comment to have me ignore it.
• Look at CI failures and help fix them

⚙️ Control Options:

• Disable automatic comment and CI monitoring

[jordanrburger]

Do we want this in here considering it will be a public repo?

[Matovidlo]

@jordanrburger That's what I was fighting a bit with. We have developer role there, we use it in our own prompts. I am open to create new role and put it there, just it will need user to switch between roles. Or We would need to create another repo with private commands that are handy for us not just our customers.
@ujovlado WDYT?

[ujovlado]

It is a showcase what we can do. I would vote for renaming to to Keboola Developer … or group somehow to "keboola internal". If it will interfere with other tools (I mean if this one will be decreasing quality of others), then we should separate it.

[ujovlado]

Otherwise I am ok to have it there. 😉