rrr

Generates session retrospectives with AI diary entries and lessons learned from git history.

Updated Feb 28, 2026
One-click install
npx skills add https://github.com/heman1033p-create/cv-v2-kongphop --skill rrr-heman1033p-create
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: rrr
Source: https://github.com/heman1033p-create/cv-v2-kongphop/tree/main/.gemini/skills/rrr
Command: npx skills add https://github.com/heman1033p-create/cv-v2-kongphop --skill rrr-heman1033p-create

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Developers often finish coding sessions without capturing what was accomplished, what went wrong, or what was learned, losing valuable context for future work. This Skill automates the creation of structured session retrospectives, preserving insights and lessons learned in a persistent memory vault. ## Core Features & Use Cases - Multiple Retrospective Modes: Run a quick default retro, a detailed full-template retro, a timeline reconstruction from session logs (--dig), or a deep 5-agent parallel analysis (--deep). - Structured Output: Generates retrospectives with timeline, files modified, first-person AI diary, honest feedback, lessons learned, and next steps, saved to a dated memory vault. - Oracle Memory Sync: Automatically syncs extracted lessons to an Oracle knowledge store with tagged concepts for future retrieval. - Use Case: At the end of a long refactoring session, invoke the retrospective command to automatically gather git history, write a reflective diary entry, document friction points, and store reusable lessons for future sessions. ## Quick Start Ask the AI to run a session retrospective by saying "rrr" or "wrap up this session" at the end of your work.

Frequently Asked Questions about rrr

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I create a session retrospective after coding?▼

Invoke the rrr command at the end of your work session. It gathers git log and diff statistics, then writes a retrospective with timeline, files modified, AI diary, honest feedback, and lessons learned to a dated memory vault path.

What retrospective modes are available for session summaries?▼

Four modes exist: default for a quick retro, --detail for a full template with metrics, --dig to reconstruct timelines from session .jsonl logs, and --deep for a 5-agent parallel analysis. Only --deep spawns subagents.

Where are retrospective files stored?▼

Retrospectives are written to ψ/memory/retrospectives/YYYY-MM/DD/ and lessons to ψ/memory/learnings/, which symlink to a shared vault at ~/.oracle. These vault files are shared state and must not be committed to git repositories.

Can I reconstruct a retrospective from a past session?▼

Yes, the --dig mode scans Claude Code session .jsonl files using the trace skill's dig instructions to rebuild a past session timeline, then writes a full retrospective using the detailed template with the reconstructed timeline table.

What are the limitations of the deep analysis mode?▼

The --deep mode is the only mode permitted to spawn subagents, launching 5 parallel agents for git analysis, file changes, timeline reconstruction, pattern extraction, and Oracle memory search. All other modes run on the main agent only.