encode-lessons

Convert recurring corrections and preventable mistakes into durable structural guards.

2|Updated May 14, 2026
One-click install
npx skills add https://github.com/AnthonyPoschen/agent-skills --skill encode-lessons-anthonyposchen
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: encode-lessons
Source: https://github.com/AnthonyPoschen/agent-skills/tree/main/skills/encode-lessons
Command: npx skills add https://github.com/AnthonyPoschen/agent-skills --skill encode-lessons-anthonyposchen

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Repeated corrections and preventable failures waste effort because reminders rely on human attention. This Skill helps you turn recurring lessons into the smallest durable mechanism—types, lint rules, CI checks, templates, or runtime validation—so the same mistake cannot happen again. ## Core Features & Use Cases - Lesson Classification: Distinguish one-off issues from recurring or costly patterns before changing the system. - Guard Selection: Choose the strongest suitable mechanism, from making invalid states impossible via types down to concise judgment-based rules. - Loop Closing: Place the guard in the owning layer, prove it catches the bad path, and remove redundant reminders. - Use Case: After a reviewer corrects the same API misuse for the third time, encode a lint rule or banned API check in CI so the pattern is blocked automatically instead of relying on code review. ## Quick Start Ask the agent to encode the recurring correction you just gave into the smallest durable guard in the codebase.

Frequently Asked Questions about encode-lessons

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

FAQPage Schema
How do I prevent the same code review comment from recurring?▼

Encode the correction as a durable guard such as a lint rule, CI check, type constraint, or canonical helper. Place it in the layer that owns the failure and verify it actually blocks the bad path.

When should I turn a mistake into a permanent rule?▼

Create a durable guard when a correction recurs or when a single mistake is high-risk or high-cost. One-off minor issues should just be fixed directly without adding permanent rules.

What kinds of guards can prevent recurring coding mistakes?▼

Options range from strongest to weakest: making invalid states impossible via types or data shape, static checks and banned APIs, canonical helpers and templates, runtime validation, and concise judgment-based rules with concrete failure examples.

When should I not encode a lesson as a structural guard?▼

Avoid guards for isolated one-off events, vague preferences without a proven pattern, or decisions that genuinely require human judgment. Building a framework for a minor preference adds cost without reducing real risk.

What should I do after adding a guard for a recurring mistake?▼

Prove the guard catches the bad path, then remove or shorten redundant reminders while keeping useful rationale and exceptions. If the guard exceeds the current task's scope, report the pattern and get approval first.