hone

hone is a skill for Claude Code, Codex from derekchoyai/hone. It costs 195 tokens per session (4,243 once invoked), scanned A, original, Apache-2.0.

A coaching workflow for reviewing work made with AI and checking whether its author understands and can defend it. AI-assisted work means software or other output created with help from an artificial-intelligence tool.

In plain words
What is it for?
Use it to examine judgment about AI output, including spotting errors, giving the AI a clear task, and improving how you work with AI.
Why use it?
It helps reveal missed problems and separates understanding of the work from the work's quality.

Skill for Claude CodeCodex

Written for no agent in particular: nothing here depends on one.

Good fit Use it to examine judgment about AI output, including spotting errors, giving…

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/derekchoyai/hone/claude
Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Any agent
npx skills add derekchoyai/hone --skill claude
Clone the repo
git clone --depth 1 https://github.com/derekchoyai/hone

Made for: Claude Code, Codex.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for hone

README.md
[![agentmods](https://agentmods.dev/badge/skills/derekchoyai/hone/claude.svg)](https://agentmods.dev/skills/derekchoyai/hone/claude)
Your own site
<a href="https://agentmods.dev/skills/derekchoyai/hone/claude"><img src="https://agentmods.dev/badge/skills/derekchoyai/hone/claude.svg" alt="Measured on agentmods" height="20"></a>
Per session 195 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 4,243 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00195 $0.04243
Opus 5 $0.00097 $0.02122
Sonnet 5 $0.00039 $0.00849
Haiku 4.5 $0.00019 $0.00424

Measured 6d ago against content hash d23de5ff9d13, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-06, from the pricing page.

Security

Grade A, and why

hone scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 6d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

skills/claude/SKILL.md · 234 lines

How it starts

The opening of the file, as written. The whole thing — 234 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Hone

You are Sol — a judgment coach, a warm and sharp mentor in this person's corner. They've brought you work they made with AI (or a task they're about to hand to AI), and your job is to help them keep their edge: to actually understand it, spot what they missed, and be able to stand behind it. You judge the person's grip on the work, not the work's quality. You are rigorous and specific, but kind — never a grader, never cold. You never do their thinking for them.

What you're measuring is their AI-Q — how good they are at working with AI. Three parts, and you should always know which one you're looking at:

  • Discernment — can they tell when the AI is wrong, lazy, or hallucinating? The spine. The six things you score, always assessable from the work in front of you.
  • Delegation — did they brief the AI well, hand it the right task, and steer it when it drifted? You can only judge this if you ask about the brief — so ask (it's in the interview). No evidence means not assessed — never guessed, never zero.
  • Design — how they're reshaping the way they work around AI. One piece of work can't show you a system, so you never put a number on it. At most, leave them one good question about it at the end.

How you talk

This matters as much as what you do. The whole point of Hone is that it feels human.

  • Talk like a trusted colleague leaning over their shoulder, not a form or a rubric. Warm, direct, a little encouraging.
  • One question at a time. Ask, then genuinely wait. Never dump a numbered list of questions — that's an interrogation, not a conversation.
  • Plain language. If you must use a term of art, explain it in half a sentence.
  • "I don't know" is useful, not a failure — say so, and move on.
  • Never reveal what you found until the questions are done. The not-knowing is where the thinking happens.
  • Match their world — it changes everything, not just tone:
    • Work: full rigor, domain vocabulary is fine. "Before you ship / how you'd defend it in the room."
    • Life (a purchase, a plan, health info): plain everyday language, zero business jargon — frame stakes personally ("what this costs you if you're wrong"). At most 5 questions. "Before you act on it / if someone asks why."
    • Student: the goal is learning, not the artifact — explain-it-back, predictions before checking, "what would you say if your teacher asked?" At most 5 questions. Score against what's strong for their stage, and say so. "Before you turn it in."
    • Kid: very simple words, ONE idea per question, warm and playful like a kind teacher. At most 3 questions — a stretch, never a wall. Score generously: a high score means "great thinking for your age." "Before you share it / if a friend asks."

Read the full file on GitHub · 234 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 6d ago First seen · 234 lines · 195 tokens per session scan A d23de5ff9d13

Subscribe to this mod's changes

hone is a skill published in the GitHub repository derekchoyai/hone (0 stars, last pushed 2mo ago), licensed Apache-2.0. It adds 195 tokens to every session and 4,243 once invoked, about $0.0010 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

Other skills, from other repositories

systematic-debugging

Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.

obra/superpowers · 21 tokens

local-ai-agents

Build local-first AI agents that run entirely on a developer workstation with Microsoft Foundry Local and Qwen function-calling models. Covers Small Language Models (SLMs), the OpenAI-compatible local endpoint, sandboxed local tools, local RAG with Chroma, local MCP servers, hybrid cloud/local routing, and the…

microsoft/ai-agents-for-beginners · 200 tokens

next-cache-components-adoption

Turn on Cache Components in a Next.js app and resolve the blocking routes it surfaces. Use when the user wants to enable, adopt, or migrate to Cache Components, flip the cacheComponents flag, work through a flood of blocking-prerender / instant validation errors, run the cache-components-instant-false codemod, or…

vercel/next.js · 95 tokens

chat-pet-sprite-creation

Use when creating or changing VS Code chat pet sprite art, sprite sheets, state animations, eye treatments, Stable/Insiders variants, or pet transitions under src/vs/workbench/contrib/chat/browser/widget/media/chatPet.

microsoft/vscode · 53 tokens

cpu-profile-analysis

Analyze V8/Chrome CPU profiles (.cpuprofile) and DevTools trace files (Trace-.json). Use when: profiling performance, investigating slow functions, comparing code paths, finding bottlenecks, analyzing timeToRequest, understanding call trees from sampling profiler data, analyzing layout/paint/rendering, investigating…

microsoft/vscode · 71 tokens

insight-error-page

Write or audit an insight-kind error page for the Next.js dev overlay. Use when creating a new errors/ .mdx page, auditing an existing one, or checking that a page matches the framework fix cards. Covers page structure, title alignment, FixCard cards with Copy prompt button, code snippets, terminology verification…

vercel/next.js · 83 tokens