learning-report

Overview

Aggregate learning logs, archived lessons, and profile data into a progress report over a given timeframe. Highlight growth, identify patterns, and suggest next steps.

When to Use

• User wants to see what they've learned recently
• Periodic self-review (weekly, biweekly, monthly)
• Preparing for a 1:1 or performance review
• Looking for article ideas based on recent learnings

Prerequisites

Profile must exist at ~/.claude/learning/profile.md. At least one session log should exist in ~/.claude/learning/logs/ for the report to be meaningful.

Process

1. Parse Timeframe

Accept an argument for the timeframe. Default: 2 weeks.

Supported formats:

• 2w — 2 weeks
• 1m — 1 month
• 3d — 3 days
• No argument — default to 2w

Calculate the start date from today minus the timeframe.

2. Read Data

• ~/.claude/learning/profile.md — gaps for comparison
• ~/.claude/learning/logs/ — all session logs within the timeframe
• ~/.claude/learning/lessons/archive/ — lessons closed within the timeframe
• ~/.claude/learning/lessons/current.md — active lesson (if any)
• ~/.claude/learning/insights.md — learning intelligence data (may not exist)

3. Generate Report

Structure the report with these sections:

Summary

• Number of sessions in the period
• Approximate total time invested (sum of duration_approx from logs)
• Lessons completed / postponed / abandoned / in progress

Topics Covered

• List each topic worked on with milestone progress

Key Learnings

• Aggregate and deduplicate learnings across all session logs
• Group by topic for readability

Recurring Struggles

• Themes that appear in "Struggled With" or "Still Fuzzy On" across multiple sessions
• These are the areas that need more attention

Gap Progress

• Compare profile gaps against lessons completed and learnings gathered
• Which gaps have been addressed? Which are untouched?

Patterns

• Abandoned/postponed patterns (if any): "You postponed 2 frontend lessons — is frontend still a priority?"
• Session frequency: "You averaged 3 sessions/week"
• Style preferences observed

Learning Intelligence (only if insights.md exists)

• Strengths confirmed during this period
• Recurring struggles with session counts
• Difficulty progression for each gap (beginner → intermediate → advanced)
• Any incidental notes logged during the period

Suggested Next Steps

• Recommend topics to revisit based on "still fuzzy" themes
• Suggest gaps to tackle next
• Optionally: propose article ideas based on what the user learned well enough to teach

Example output:

# Learning Report — Mar 17 to Mar 31, 2026

## Summary
- **Sessions:** 6 sessions over 2 weeks
- **Time invested:** ~9 hours
- **Lessons:** 1 completed, 1 in progress

## Topics Covered
- React Dashboard (completed) — 4/4 milestones
- GraphQL API Design (in progress) — 1/3 milestones

## Key Learnings
**React:**
- Components are functions returning JSX; state drives re-renders
- useEffect cleanup prevents memory leaks on unmount

**GraphQL:**
- Schema-first design prevents client/server drift

## Recurring Struggles
- State management patterns (came up in 3 sessions)
- Async error handling (2 sessions)

## Gap Progress
- Frontend: addressed (React dashboard completed)
- Backend APIs: in progress (GraphQL)
- Infrastructure: untouched

## Learning Intelligence
- **Confirmed strengths:** LCEL chain composition, prompt templating
- **Recurring struggles:** async error handling (4 sessions), state management (3 sessions)
- **Difficulty progression:** backend APIs: intermediate → advanced, AI/orchestration: beginner → intermediate

## Suggested Next Steps
- Revisit state management — still fuzzy after multiple sessions
- Continue GraphQL lesson (1 milestone left this week)
- Consider starting an infrastructure topic next

4. Save Report

Write to ~/.claude/learning/reports/YYYY-MM-DD-<period>.md

• Example: 2026-03-31-2w.md

5. Display

Show the full report to the user in the conversation. Mention where it was saved.

Important

• If no logs exist for the timeframe, tell the user and suggest a different range
• Deduplicate learnings — the same insight may appear in multiple session logs
• The "article ideas" section is optional — only include if the user has substantial learnings
• Keep the tone encouraging — highlight growth, not just gaps