AGENT SKILL

principle-build-the-lever

Build the smallest rerunnable tool that performs or proves non-trivial work. Use for migrations, repetitive edits, analyses, audits, generators, or delegated campaigns where hand work would drift or be hard to review.

License 许可
MIT
所属仓库
caezium/skills
最近更新
2026-08-24
Stars
★ 3
Forks
0

01 前言 Front Matter

name
principle-build-the-lever
description
Build the smallest rerunnable tool that performs or proves non-trivial work. Use for migrations, repetitive edits, analyses, audits, generators, or delegated campaigns where hand work would drift or be hard to review.

02 资源文件 1

03 SKILL.md 正文

Build the lever

When work is non-trivial, make the method a rerunnable artifact: a codemod, script, query, generator, checker, or delegate contract. The lever buys both throughput and reviewability.

Pattern

  1. Perform one representative unit by hand to learn the recipe.
  2. Build the smallest deterministic tool that reproduces or verifies that unit.
  3. Diff the tool's result against the hand-worked unit.
  4. Make reruns safe: idempotent where possible, scoped, and loud on partial failure.
  5. Run it across the remaining units and retain its proof output.

Prefer a deterministic script to agents hand-applying the same transformation. When the work genuinely requires judgment and delegation is authorized, make the lever a read-only contract every worker follows: goal, recipe, verification predicate, output schema, and do-not-touch fences. Keep that contract outside worker write scopes.

Applying this principle produces a file. If no script, generator, query, checker, or delegate contract exists, the lever was not built. Skip only a couple of obvious edits whose result is cheaper to inspect than the tool would be to maintain.

Commit the lever when the task or its verification will recur; otherwise keep it as a visible run artifact.

<!-- Adapted from poteto/pstack principle-build-the-lever, MIT License, Lauren Tan, cursor/plugins@51a96e0. -->

04 同仓库更多 Skills

caezium
caezium/skills

writing-for-agents

Writing documents for agents. Use when creating or editing skills, or modifying AGENTS.md or CLAUDE.md.

内容创作者写作与文案
★ 3
caezium
caezium/skills

wrangler

Cloudflare Workers CLI for deploying, developing, and managing Workers, KV, R2, D1, Vectorize, Hyperdrive, Workers AI, Containers, Queues, Workflows, Pipelines, and Secrets Store. Load before running wrangler commands to ensure correct syntax and best practices. Biases towards retrieval from Cloudflare docs over pre-trained knowledge.

后端与接口
★ 3
caezium
caezium/skills

workers-best-practices

Reviews and authors Cloudflare Workers code against production best practices. Load when writing new Workers, reviewing Worker code, configuring wrangler.jsonc, or checking for common Workers anti-patterns (streaming, floating promises, global state, secrets, bindings, observability). Biases towards retrieval from Cloudflare docs over pre-trained knowledge.

★ 3
caezium
caezium/skills

wizard

Generate an interactive bash wizard that walks a human through steps only they can perform. Use when provisioning infrastructure, setting up credentials or CI secrets, walking an unfamiliar third-party dashboard, or running a one-off migration or cutover. Don't invoke this for steps the agent can perform itself.

★ 3
caezium
caezium/skills

why

Use for 'why does X work this way', 'why we picked Y', design rationale, regressions, postmortems, or data-backed thresholds. Discovers available evidence sources, searches each relevant category, and returns a cited read on decisions and tradeoffs. Use how for runtime behavior.

★ 3
caezium
caezium/skills

week-prep

Generate a briefing with calendar, tasks, and attention flags. Use for "week prep", "briefing", or "what's on this week".

Skill与智能体开发
★ 3