mirror of
https://github.com/microsoft/SkillOpt.git
synced 2026-07-03 14:02:58 +08:00
Bug fixes: - #52: bundle run-sleep.sh in Claude Code plugin + 4-level fallback - #58: add skillopt-sleep console script entry point in pyproject.toml - #62: filter headless claude -p replay sessions from harvest Plugin sync (Claude Code / Codex / Copilot / OpenClaw): - Document all 22 CLI flags, 7 actions, 4 backends across all SKILL.md files - Document config keys (preferences, gate_mode, dream_rollouts, etc.) - Document memory consolidation (evolve_memory / evolve_skill) - Add schedule/unschedule to all plugins - Copilot MCP: expand schema from 3 → 16 params + schedule tools - OpenClaw: add schedule/unschedule subcommands via shared scheduler Tests: - Cross-plugin parity test (prevents future feature drift) - MCP schema completeness test Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2.0 KiB
2.0 KiB
SkillOpt-Sleep (offline self-evolution)
This project has SkillOpt-Sleep available via an MCP server (skillopt-sleep).
It gives the agent a nightly "sleep cycle": it reviews past sessions, replays
recurring tasks offline, and consolidates validated memory + skills behind a
held-out gate.
When the user asks to "run the sleep cycle", "review my past sessions", "learn my preferences", or "make the agent improve from past usage", use the MCP tools:
sleep_status— what's happened + the latest staged proposalsleep_dry_run— safe preview, stages nothingsleep_run— full cycle, stages a reviewed proposal (nothing live changes)sleep_adopt— apply the staged proposal (backs up first)sleep_harvest— list mined recurring taskssleep_schedule— install a nightly cron entry (sethour/minute)sleep_unschedule— remove the nightly cron entry
Key parameters (pass as MCP tool arguments)
backend—mock(default, free),claude,codex, orcopilotsource—claude,codex, orauto(where to read transcripts)target_skill_path— explicit SKILL.md to evolvetasks_file— pre-built TaskRecord JSON (skip harvest)max_tasks/max_sessions— cap workloadauto_adopt— auto-adopt if the gate passesjson— machine-readable output for programmatic use
Advanced config (~/.skillopt-sleep/config.json)
preferences— free-text house rules for the optimizergate_mode—on(default) oroff;dream_rollouts— >1 for more signalevolve_memory/evolve_skill— toggle which docs consolidate
Always show the user the held-out baseline → candidate score and the proposed
edits before suggesting sleep_adopt. Never hand-edit the user's memory/skill
files; only sleep_adopt does that, with a backup.