/daily-standup

name: daily-standup
description: >
  Generate a daily standup summary from recent Claude Code sessions — completed
  work grouped by project (cwd), session costs from the pricing engine,
  tool invocations, error/compaction/APIError events, and turn velocity
  metrics from session metadata (turn_count, total_turn_duration_ms).

Daily Standup

Generate a daily standup report from Claude Code Agent Monitor data.

Input

The user provides: **$ARGUMENTS**

This may be:

• "today" or empty (default: last 24 hours)
• "yesterday" for the previous day
• A specific date: "2025-04-10"

Procedure

1. **Fetch recent session data** from `http://localhost:4820`:

• `GET /api/sessions?limit=50` (default sort: most recently updated first)
• Filter sessions that started within the target day
• For each matching session: `GET /api/events?session_id={session_id}`

2. **Compile standup sections**:

✅ What I accomplished

• List each completed session with:
• Brief description (from session name or first tool's context)
• Working directory (project context)
• Key tools used and outcomes
• Duration and model used
• Group by project/working directory if multiple

⚠️ Issues encountered

• Sessions that ended in `error` or `abandoned` status
• Tools that failed (from error events)
• Compaction events (hit context limits)
• Unusually long sessions (>2x average duration)

📋 Key metrics

• Total sessions: N
• Total time spent: X hours Y minutes
• Tools invoked: N (top 3 listed)
• Estimated cost: $X.XX
• Completion rate: N%

🔮 Suggested focus areas

• Based on incomplete/error sessions, suggest what to revisit
• Based on tool patterns, suggest workflow improvements

3. **Format for standup**:

• Keep it concise — aim for a 2-minute read
• Lead with accomplishments
• Be honest about blockers
• Make metrics scannable

Output Format

Present as a clean standup report with emoji section headers, bullet points for items, and a compact metrics table. Add a one-line summary at the top suitable for pasting into Slack or a team channel.