メニュー
Create session retrospective with AI diary and lessons learned. Use when user says "rrr", "retrospective", "wrap up session", "session summary", or at end of work session.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
| name | rrr |
| description | Create session retrospective with AI diary and lessons learned. Use when user says "rrr", "retrospective", "wrap up session", "session summary", or at end of work session. |
| argument-hint | [--detail | --dig | --deep] |
"Reflect to grow, document to remember."
/rrr # Quick retro, main agent
/rrr --detail # Full template, main agent
/rrr --dig # Reconstruct past timeline from session .jsonl
/rrr --deep # 5 parallel agents (read DEEP.md)
NEVER spawn subagents or use the Task tool. Only --deep may use subagents.
/rrr, /rrr --detail, and /rrr --dig = main agent only. Zero subagents. Zero Task calls.
date "+%H:%M %Z (%A %d %B %Y)"
git log --oneline -10 && git diff --stat HEAD~5
ENCODED_PWD=$(pwd | sed 's|^/|-|; s|/|-|g')
PROJECT_DIR="$HOME/.claude/projects/${ENCODED_PWD}"
LATEST_JSONL=$(ls -t "$PROJECT_DIR"/*.jsonl 2>/dev/null | head -1)
if [ -n "$LATEST_JSONL" ]; then
SESSION_ID=$(basename "$LATEST_JSONL" .jsonl)
echo "SESSION: ${SESSION_ID:0:8}"
fi
If detected, include in retrospective header:
📡 Session: 74c32f34 | repo-name | Xh XXm
If detection fails, skip silently.
Path: ψ/memory/retrospectives/YYYY-MM/DD/HH.MM_slug.md
mkdir -p "ψ/memory/retrospectives/$(date +%Y-%m/%d)"
Write immediately, no prompts. Include:
Path: ψ/memory/learnings/YYYY-MM-DD_slug.md
arra_learn({ pattern: [lesson content], concepts: [tags], source: "rrr: REPO" })
Retro files are written to vault (wherever ψ symlink resolves).
Do NOT git add ψ/ — it's a symlink to the vault. Vault files are shared state, not committed to repos.
Same flow as default but use full template:
# Session Retrospective
**Session Date**: YYYY-MM-DD
**Start/End**: HH:MM - HH:MM GMT+7
**Duration**: ~X min
**Focus**: [description]
**Type**: [Feature | Bug Fix | Research | Refactoring]
## Session Summary
## Timeline
## Files Modified
## Key Code Changes
## Architecture Decisions
## AI Diary (150+ words, vulnerable, first-person)
## What Went Well
## What Could Improve
## Blockers & Resolutions
## Honest Feedback (100+ words, 3 friction points)
## Lessons Learned
## Next Steps
## Metrics (commits, files, lines)
Then steps 3-5 same as default.
Retrospective powered by session goldminer. No subagents.
Discover project dirs using full-path encoding (same as Claude's .claude/projects/ naming), including worktree dirs:
ENCODED_PWD=$(pwd | sed 's|^/|-|; s|/|-|g')
PROJECT_BASE=$(ls -d "$HOME/.claude/projects/${ENCODED_PWD}" 2>/dev/null | head -1)
export PROJECT_DIRS="$PROJECT_BASE"
for wt in "${PROJECT_BASE}"-wt*; do [ -d "$wt" ] && export PROJECT_DIRS="$PROJECT_DIRS:$wt"; done
Then run dig.py to get session JSON:
python3 ~/.claude/skills/dig/scripts/dig.py 0
Also gather git context:
date "+%H:%M %Z (%A %d %B %Y)"
git log --oneline -10 && git diff --stat HEAD~5
Use the session timeline data to write a full retrospective using the --detail template. Add the Past Session Timeline table after Session Summary, before Timeline.
Write lesson learned, oracle sync.
Do NOT git add ψ/ — vault files are shared state, not committed to repos.
Read DEEP.md in this skill directory. Only mode that uses subagents.
If the Oracle was born via /awaken wizard v2, CLAUDE.md may contain:
auto, /rrr runs are expected and welcomed. If manual, only run when explicitly asked.Check CLAUDE.md for these fields. If not present, use defaults (auto memory, standard depth).
--deep may)Ask another agent a question and get an answer in real-time. Uses Agent Pool delegation API (sync mode). Auto-selects the right agent if not specified. Use when you need information, opinions, or validation from another agent NOW — not async threads.
Fast agent query via direct tmux (~5-10s). No audit trail. Use for quick checks, yes/no answers, and one-liners where /ask (30-40s) is too slow.
Delegate QA verification to another agent (async). Uses issue as completion artifact — no sync timeout. Use for long-running QA, review, or investigation tasks where /ask would time out.
Multi-agent brainstorm with refinement loop. Agents brainstorm, critique, and refine ideas together via delegation API. Supports 3 patterns: fan-out, round-robin, creative-critic. Live Telegram updates + markdown archive + GitHub Discussion as live thread (--publish).
Query schedule via Oracle API (Drizzle DB). Use when user says "schedule", "upcoming events", "what's on today", "calendar".
Guided Oracle birth and awakening ritual. Default is Full Soul Sync (~20min), or --fast (~5min). Use when creating a new Oracle in a fresh repo, when user says "awaken", "birth oracle", "create oracle", "new oracle", or wants to set up Oracle identity in an empty repository. Do NOT trigger for general repo setup, git init, or project scaffolding without Oracle context.