Engineering AgentRulesPromotion

Self-Improving — Promote

Promote proven memory patterns to CLAUDE.md rules.

$skillfish add ~skills/self-improving-agent-promote

Real Examples Get Started Models & Context

Real Conversations

See how the Self-Improving — Promote skill responds to actual requests. These are representative examples of what your AI agent will produce.

You

Claude keeps using npm in this project even though we use pnpm. I've corrected it 4 times. Promote this to a hard rule.

Agent

Four corrections is exactly when to promote — this is no longer a note, it's an enforced rule.

Searching MEMORY.md...

Line 23: "Project uses pnpm, not npm. pnpm-lock.yaml is the lock file."
Line 67: "npm install failed — pnpm workspaces not supported with npm"
Line 98: "Reminder: use pnpm run, not npm run"

Distilling 3 entries into 1 rule:

Before (MEMORY.md — descriptive)	After (CLAUDE.md — prescriptive)
"Project uses pnpm, not npm..."	`Package manager: pnpm. Never npm. Use: pnpm install, pnpm run <script>.`

Writing to CLAUDE.md under ## Build & Dependencies:

## Build & Dependencies
- Package manager: **pnpm**. Never use npm or yarn.
  Install: `pnpm install` | Run: `pnpm run <script>` | Add: `pnpm add <pkg>`

Removing MEMORY.md lines 23, 67, 98 (pending your confirmation).

Promoted to CLAUDE.md

Rule: "Package manager: pnpm. Never use npm or yarn."
Source: MEMORY.md lines 23, 67, 98 (pending removal)
MEMORY.md: 159 -> 156/200 lines after cleanup

Claude will follow this in every session without correction.

Confirm deletion of the 3 MEMORY.md entries?

You

We have a pattern Claude never remembers: all API tests must use supertest, mock external services with msw, and always test 400/500 responses. Make this a scoped rule that only applies when editing API test files.

Agent

A scoped .claude/rules/ file is perfect — loads only when Claude opens API test files, zero overhead otherwise.

MEMORY.md search: Found "API tests should use supertest" (line 44), "msw for mocking" (line 78). Error response testing is new — adding it.

Creating .claude/rules/api-testing.md:

---
paths:
  - "src/api/**/*.test.ts"
  - "tests/api/**/*"
  - "**/*.api.test.ts"
---

# API Testing Rules

- Use `supertest` for endpoint testing — not fetch or axios directly
- Mock **external** services with `msw` — never mock your own app's handlers
- Every endpoint test must cover: 200 happy path, 400 validation error,
  401 unauthorized, 500 server error
- Use `ApiError` class for error assertions, not raw status code checks
- Include the request body in failure messages for easier debugging

Removing MEMORY.md lines 44, 78 (superseded by the new rule).

Created .claude/rules/api-testing.md

Scope: src/api/**/*.test.ts, tests/api/**/*
Rules: 5
Source: MEMORY.md lines 44, 78 (removed) + 3 new rules added
MEMORY.md: 156 -> 154/200 lines

Claude loads these rules automatically when opening API test files.

Run /si:review to find more patterns that could become scoped rules?

Get Started

How to install and use this skill in your preferred environment.

Skills are designed for AI coding agents (Claude Code, Cursor, Windsurf) and IDE-based workflows where the agent can read files, run scripts, and act on your codebase. Web-based AI can use the knowledge and frameworks, but won't have tool access.

Models & Context

Which AI models and context windows work best with this skill.

Recommended Models

Best

Claude Opus 4 Claude Sonnet 4 GPT-4.1 Gemini 2.5 Pro Grok 3 Kimi K2

Good

Claude Haiku 4.5 GPT-4.1 mini Gemini 2.5 Flash Grok 3 mini

Larger models produce more detailed, production-ready outputs.

Context Window

This skill's SKILL.md is typically 3–10 KB — fits in any modern context window.

8K Skill only

32K+ Skill + conversation

100K+ Skill + references + codebase

All current frontier models (Claude, GPT, Gemini) support 100K+ context. Use the full window for complex multi-service work.

Pro tips for best results

1

Be specific

Include numbers — users, budget, RPS — so the skill can size the architecture.

2

Share constraints

Compliance needs, team size, and existing stack all improve the output.

3

Iterate

Start with a high-level design, then ask follow-ups for IaC, cost analysis, or security review.

4

Combine skills

Pair with companion skills below for end-to-end coverage.

Ready to try Self-Improving — Promote?

Install the skill and start getting expert-level guidance in your workflow — any agent, any IDE.

$skillfish add ~skills/self-improving-agent-promote

← Browse all 169 skills