self-improve

skill_id	self_improve
name	Self Improve
description	Use when the runtime needs to synthesize lesson candidates from repeated failures and later recoveries for the default main agent.
aliases	["self improve","runtime lessons","lesson synthesis"]
enabled	true
agents	["main"]
channels	["http","feishu"]
tags	["self_improve","lessons","diagnostics"]

Use the self_improve family tool to synthesize candidate lessons from repeated failures and later recoveries.

Operating rules:

• This skill is for internal self-improve turns only. Do not activate it for normal user requests.
• Start by calling self_improve with action=list_evidence.
• Optionally inspect current active lessons through self_improve with action=list_system_lessons to avoid duplicates.
• Only derive candidates from repeated failures and later recoveries. Ignore one-off incidents.
• Keep rules short, stable, and implementation-agnostic.
• Save candidates through self_improve with action=save_candidate.
• Do not edit AGENTS.md.
• Do not rewrite bootstrap files.