skillopt-sleep

Consolidates past Claude sessions into gated CLAUDE.md and SKILL.md improvements.

Updated Aug 14, 2026
One-click install
npx skills add https://github.com/Rohithdgrr/REEK-uninstaller --skill skillopt-sleep-rohithdgrr
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: skillopt-sleep
Source: https://github.com/Rohithdgrr/REEK-uninstaller/tree/main/.opencode/skills/skillopt-sleep
Command: npx skills add https://github.com/Rohithdgrr/REEK-uninstaller --skill skillopt-sleep-rohithdgrr

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve? Your agent forgets everything between sessions, so recurring mistakes and preferences never turn into lasting improvements. This Skill runs an offline sleep cycle that reviews past Claude Code sessions, replays recurring tasks, and consolidates validated learnings into memory and skill files. ## Core Features & Use Cases - Six-stage sleep cycle: harvest past session transcripts, mine recurring tasks, replay them offline, consolidate bounded edits, stage proposals, and adopt only after a held-out validation gate passes. - Gated memory and skill evolution: proposes bounded add/delete/replace edits to CLAUDE.md and SKILL.md, accepting changes only when they strictly improve held-out scores, with backups before adoption. - Scheduling and safe preview: install a nightly cron entry, run a zero-cost mock backend dry-run, or spend real API budget with claude/codex backends. - Use Case: Schedule a nightly run so your agent reviews the day's sessions at 3 AM, stages a proposed CLAUDE.md update encoding your preferences, and shows you the held-out score improvement before you adopt it. ## Quick Start Ask your agent to run the skillopt-sleep dry-run on the current project to preview what a nightly self-improvement cycle would propose.

Frequently Asked Questions about skillopt-sleep

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

FAQPage Schema
How do I make my Claude agent learn from past sessions?▼

Run the sleep cycle, which harvests past Claude Code session transcripts, mines recurring tasks, replays them offline, and consolidates validated learnings into CLAUDE.md and SKILL.md. Changes are only adopted after passing a held-out validation gate.

How do I schedule a nightly self-improvement run for my agent?▼

Use the schedule action of the bundled sleep.sh script with a project path, hour, and minute to install a nightly cron entry. The unschedule action removes managed entries when you no longer want automatic runs.

Does the sleep cycle spend API credits?▼

No by default. The mock backend is deterministic and spends nothing, making it safe for testing the plumbing. Add the claude or codex backend flag only when you want genuine replay using your real API budget.

Can the sleep cycle modify my CLAUDE.md without permission?▼

No. The cycle only stages proposed files, diffs, and a report in a staging directory. Live CLAUDE.md and SKILL.md files change only through the explicit adopt action, which creates backups first.

What happens if a proposed edit makes the agent worse?▼

The held-out validation gate rejects any edit that does not strictly improve the score, and rejected edits become negative feedback. A deterministic demo experiment proves the gate blocks regressions without API spend.