digest

name	digest
description	Summarize vault learnings and activity for today or a date range.
user-invocable	true
argument-hint	[today | YYYY-MM-DD | YYYY-MM-DD..YYYY-MM-DD]
context	fork
agent	researcher
allowed-tools	["Read","Glob","Grep"]
model	sonnet

Learning Digest

Summarize vault learnings and activity for a given time period.

Current State

!ls vault/learnings/ 2>/dev/null | wc -l | tr -d ' ' total learning files !ls vault/learnings/*-$(date +%Y%m%d)-* 2>/dev/null || echo "(none today)" !cat vault/daily/$(date +%Y-%m-%d).md 2>/dev/null || echo "(no daily digest yet)"

Date Parsing

Parse $ARGUMENTS to determine the date range:

• No args or today — today's date only
• YYYY-MM-DD — specific date
• YYYY-MM-DD..YYYY-MM-DD — date range (inclusive)

Convert dates to the YYYYMMDD format used in filenames.

Analysis Steps

1. Find matching files: Glob for vault/learnings/*-YYYYMMDD-* for each date in range
2. Read each file: Extract frontmatter (type, status, priority/severity, tags, pattern-key, recurrence-count)
3. Aggregate:
   • Count by type (LRN/ERR/FEAT)
   • Count by status
   • Count by area
   • List any with recurrence-count >= 2 (approaching promotion)
   • List any FEAT entries with status requested (potential quick wins)
4. Check for daily digest: Read vault/daily/YYYY-MM-DD.md if it exists
5. Compile summary

Output Format

## Learning Digest: YYYY-MM-DD

### Activity
- 3 new learnings, 1 error, 0 feature requests
- 2 entries resolved today
- 1 pattern approaching promotion threshold

### New Entries
| ID | Type | Title | Status |
|----|------|-------|--------|
| LRN-20250310-001 | learning | Title here | new |
| ERR-20250310-001 | error | Title here | resolved |

### Patterns to Watch
- **env-var-stripping** (recurrence: 2/3 for promotion) — Claude CLI env issues
- **file-path-quoting** (recurrence: 2/3 for promotion) — Spaces in paths

### Quick Wins
- FEAT-20250310-001: "Add vault search from CLI" (complexity: simple)

### Daily Digest File
<contents of vault/daily/YYYY-MM-DD.md if it exists, or "(not yet generated)">

For date ranges, group entries by date and show per-day breakdowns followed by a range summary.