skills.rest
etylsarin
etylsarinCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

self-improvement

Manage lessons-learned knowledge for multi-agent teams via CLI.

59|12|Updated Feb 26, 2026
Search Tags:#cli#self-improvement#agent#governance#knowledge-base#lessons-learned
One-click install
npx skills add https://github.com/etylsarin/opencastle --skill self-improvement-etylsarin
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: self-improvement
Source: https://github.com/etylsarin/opencastle/tree/main/src/orchestrator/skills/self-improvement
Command: npx skills add https://github.com/etylsarin/opencastle --skill self-improvement-etylsarin

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill ensures multi-agent teams do not repeat the same mistakes by providing a structured protocol to read prior lessons before work and to write concise lessons after retries succeed, preserving institutional knowledge across sessions.

Core Features & Use Cases

  • Read-before-work enforcement: Agents consult the centralized LESSONS-LEARNED.md before performing tasks to avoid known pitfalls.
  • CLI-driven lesson creation: Agents use the opencastle lesson command to add atomic, categorized lessons after successful retries, ensuring consistent formatting and indexing.
  • Governance and quality standards: Enforces categories, severity levels, and quality criteria so lessons are actionable and discoverable.
  • Use Case: When an agent encounters a flaky build step, it retries with a new approach and immediately documents the working workaround to prevent future agents from repeating wasted effort.

Quick Start

Use the opencastle lesson CLI to record a concise lessons-learned entry immediately after a successful retry.

Frequently Asked Questions about self-improvement

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

FAQPage Schema
How do I capture lessons learned for autonomous agent teams to avoid repeated mistakes?▼

To capture lessons learned for autonomous agent teams, use a CLI-driven protocol that agents invoke after successful retries. This enforces read-before-work and write-on-retry governance via a centralized knowledge base, preventing repeated mistakes across code repository and CI sessions.

What is read-before-work governance for multi-agent code repositories?▼

Read-before-work governance for multi-agent code repositories requires agents to consult a centralized LESSONS-LEARNED.md file before performing any task. This mechanism ensures autonomous agents review prior categorized lessons and severities, actively avoiding known pitfalls during execution.

How do I format lessons learned entries for autonomous agents in CI?▼

To format lessons learned entries for autonomous agents in CI, use the CLI lesson command to append atomic, categorized entries with defined severity levels. This enforces quality standards so recorded lessons remain actionable and discoverable in the centralized file.

Can I use CLI-driven lesson creation to manage institutional knowledge for multi-agent teams?▼

Yes, you can use CLI-driven lesson creation to manage institutional knowledge for multi-agent teams. Agents execute the lesson command after successful retries, enforcing structured storage and indexing so documented workarounds persist across sessions without manual intervention.

What is the best way to document flaky build workarounds for autonomous agents?▼

The best way to document flaky build workarounds for autonomous agents is to immediately record the working approach via CLI after a successful retry. This enforces structured knowledge base creation, ensuring future agents consult the documented workaround before attempting the build step.

Related Skills

Explore More Software Engineering Skills→
cdrgurucdrguru
1

gemini-cli-init

Initialize Gemini CLI configuration and auditor persona for the PACK workflow.

Community
Intermediate
saintgo7saintgo7

dev-log

Create daily development log entries in Markdown files.

Community
Basic
drt0927drt0927
5

learn

Aggregate team learnings into a searchable append-only log with export.

Community
Advanced
KubeHealKubeHeal

github-issue-resolver

Diagnose and fix GitHub Actions failures, PR conflicts, and Dependabot PRs via GitHub CLI.

Official
Advanced