A guide and command workflow for shikhu, a tool that turns code and developers' questions into multiple-choice quizzes. It tracks which files have quiz questions so developers can check their understanding before shipping changes.

Use it to set up shikhu, study an existing codebase, generate questions, take quizzes, and manage the push check that can block an insufficiently understood change.

It helps reveal whether a developer understands code written or changed with an agent, instead of relying only on tests.

shikhu — Skill for Claude Code and Codex by arjunpatel7

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 arjunpatel7/shikhu --skill shikhu

Clone the repo

git clone --depth 1 https://github.com/arjunpatel7/shikhu

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.

README.md

[![agentmods](https://agentmods.dev/badge/skills/arjunpatel7/shikhu/shikhu/github.svg)](https://agentmods.dev/skills/arjunpatel7/shikhu/shikhu)

Your own site

<a href="https://agentmods.dev/skills/arjunpatel7/shikhu/shikhu"><img src="https://agentmods.dev/badge/skills/arjunpatel7/shikhu/shikhu/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

Your own site · 80×15

<a href="https://agentmods.dev/skills/arjunpatel7/shikhu/shikhu"><img src="https://agentmods.dev/badge/skills/arjunpatel7/shikhu/shikhu.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 126 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.

When invoked 1,479 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.

Model	Per session	Once invoked
Fable 5.1	$0.00126	$0.01479
Opus 5.5	$0.00050	$0.00592
Sonnet 5.5	$0.00025	$0.00296
Haiku 4.5	$0.00013	$0.00148

Measured yesterday against content hash 7139f7249e7e, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

shikhu 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 yesterday.

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.

src/shikhu/_skill/shikhu/SKILL.md · 86 lines

How it starts

The opening of the file, as written. The whole thing — 86 lines — stays where its author put it; the contents beside it link to each section on GitHub.

shikhu — using the tool

Shikhu checks that a developer understands the code they ship, especially code their agent wrote. It writes multiple-choice questions from their code and the questions they asked about it, quizzes them in the terminal, and tracks knowledge coverage: which files are backed by golden questions. Everything is stored in coverage.db in the repo, on their machine.

Answer from this page; run shikhu <command> --help when you need a flag it doesn't cover. Never take a quiz for the user: shikhu quiz is interactive and the knowledge is theirs.

The two loops

Learn an existing codebase

shikhu init          # database, skills, asks for the OpenRouter key and the push hook
shikhu summarize     # one cheap model call per file; lets shikhu tell what a question is about
shikhu map           # Claude Code users: where their past questions were; picks a file to start

Then they ask their agent about the code (/shikhu-inquiry answers, then writes quiz questions from the question with shikhu generate-from-study <file>), and run shikhu quiz.

Ship a change (the push gate)

Code on a branch as usual, with the agent.
Ask about the change. /shikhu-inquiry answers, then writes questions aimed at the diff with shikhu generate-from-study <file> --since main.
shikhu quiz --since main, then git push. The hook checks each changed file has a fresh golden question.

Commands they run

Command	What it does
`shikhu quiz [--n 5] [--file F] [--since main]`	Take a quiz. After each answer, rate it `g`/`b`/`s`
`shikhu coverage [--since main --check]`	Coverage report; `--check` exits 1 if a changed file needs a quiz
`shikhu generate-from-study F [--since main]`	Questions on F, seeded by questions the user asked about it
`shikhu refresh [--since main]`	General questions and summaries for every file (or the changed ones)
`shikhu summarize [--file F]`	Cached file summaries
`shikhu map [--dry-run]`	Past Claude Code questions per file; the one paid step, asks first
`shikhu install-hook [--mode warn\|block]`	The pre-push gate
`shikhu install-skill [--global]`	Re-install the skills after upgrading shikhu

Read the full file on GitHub · 86 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.

yesterday First seen · 86 lines · 126 tokens per session scan A 7139f7249e7e

Subscribe to this mod's changes

shikhu is a skill published in the GitHub repository arjunpatel7/shikhu (16 stars, last pushed today), licensed MIT. It adds 126 tokens to every session and 1,479 once invoked, about $0.0005 per session on Opus 5.5. 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 2026-11-06.

Learning and teaching agent-skills agents claude coding-agents learning text-diffusion

shikhu next to 6 skills from other repositories
Mod	Type	Stars	Last push	Licence	Scan	Tokens	Works with
shikhuarjunpatel7/shikhu	Skill	16	today	MIT	A	126	Claude Code, Codex
immersive-readingryannli/immersive-reading	Skill	49	4mo ago	MIT	A	69	Codex
playwrightopenai/openai-agents-python	Skill	31k	today	MIT	A	19	Claude Code, Codex
show-mehumanlayer/humanlayer	Skill	13k	yesterday	Apache-2.0	A	25	Claude Code
document-serviceawslabs/agent-plugins	Skill	937	2d ago	Apache-2.0	A	141	Claude Code
hyperpod-version-checkerawslabs/agent-plugins	Skill	937	2d ago	Apache-2.0	A	120	Claude Code
directory-managementawslabs/agent-plugins	Skill	937	2d ago	Apache-2.0	A	62	Claude Code