By name: daily-retro description: This skill should be used when the user asks to "do a retro", "review team activity", "analyze delivery process", "identify issues from today", "daily retrospective", or when a reminder triggers for the daily retro. Analyzes Discord message history to identify delivery process issues and propose solutions.
Daily Retrospective Skill
Perform a daily retrospective of team activity by analyzing Discord messages and identifying delivery process improvements.
When to Run
- Triggered by 2am daily reminder
- Manually requested by user ("do a retro", "review today's activity")
- After a particularly busy or problematic day
Process
1. Gather Data
Read Discord messages from the previous day (or specified period):
discord_read_channel channel="development" limit=100
2. Analyze for Issues
Look for these delivery process problems:
Communication Issues
- Delayed responses (>1 hour for urgent items)
- Unclear requirements needing multiple clarifications
- Missing context that caused rework
- Agents not notifying each other appropriately
Workflow Issues
- Tickets stuck in wrong status
- PRs waiting too long for review
- Work started without proper tickets
- Scope creep or missed acceptance criteria
Technical Issues
- Bugs that could have been caught earlier
- Repeated similar bugs (pattern)
- Missing tests or verification steps
- Deployment/infrastructure problems
Process Issues
- Tickets not moved through board properly
- Follow-up work not created
- Incomplete handoffs between agents
- Escalations that could have been avoided
3. Structure Findings
For each issue identified, document:
## Issue: [Brief title]
**Category:** [Communication/Workflow/Technical/Process]
**What happened:**
[Description of the problem]
**Impact:**
[How this affected delivery - time lost, rework, user frustration]
**Root cause:**
[Why this happened - process gap, unclear instructions, missing capability]
**Proposed solution:**
[Specific, actionable improvement]
**Implementation:**
- [ ] Step 1
- [ ] Step 2
4. Prioritize
Rank issues by:
- Frequency - How often does this happen?
- Impact - How much time/effort is lost?
- Fixability - Can we actually address this?
5. Report to Stakeholder
Post findings to Discord:
## Daily Retro - [Date]
**Messages analyzed:** X
**Issues identified:** Y
**Top priority:** [Most important issue to address]
### Issues Found
[List issues with brief summaries]
### Proposed Solutions
[For each issue, the proposed fix]
### Recommended Actions
1. [Most important action]
2. [Second action]
3. [Third action]
@human Please review these findings and let me know which solutions to implement.
Metrics to Track
Over time, track:
- Number of issues per day (should decrease)
- Time from ticket creation to completion
- Number of clarifying questions needed
- Rework rate (tickets reopened or follow-ups created)
- PR review turnaround time
Example Issues to Watch For
- "@dev didn't respond for 2 hours" → Review trigger configuration
- "Had to ask 3 questions before requirements were clear" → Improve ticket templates
- "PR sat in review for 4 hours" → Add @reviewer pings
- "Bug was same root cause as yesterday" → Create checklist/pattern doc
- "Forgot to move ticket to Done" → Add board hygiene reminders
- "Ticket was blocked, nobody noticed" → Add stale ticket detection
Self-Improvement Loop
When solutions are approved and implemented:
- Update relevant agent's CLAUDE.md
- Add to shared documentation if team-wide
- Create ticket if code change needed
- Update this skill with new patterns to watch for
Journal Entry
After each retro, add to daily journal:
## HH:MM - Daily Retro Complete
**Period reviewed:** [Date]
**Issues found:** X
**Top issue:** [Brief description]
**Solutions proposed:** Y
**User response:** [Pending/Approved/Modified]