What is prompt-improver CLAUDE.md?

A set of repository instructions for a tool that turns vague coding requests into structured XML instructions for another coding assistant to follow.

Use it when working on the prompt-improver project, running its tests, checking example prompts, or inspecting how prompts and context are assembled.

It gives developers a repeatable way to clarify incomplete requests and check the generated instructions with included Bash scripts.

How owenob1/prompt-improver configures Claude Code

Reuse

Borrowing it

Nothing to install: this file belongs to owenob1/prompt-improver. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file

curl -O https://raw.githubusercontent.com/owenob1/prompt-improver/main/CLAUDE.md

Clone the repo

git clone --depth 1 https://github.com/owenob1/prompt-improver

Made for: Claude Code.

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/instructions/owenob1/prompt-improver/claude-md/github.svg)](https://agentmods.dev/instructions/owenob1/prompt-improver/claude-md)

Your own site

<a href="https://agentmods.dev/instructions/owenob1/prompt-improver/claude-md"><img src="https://agentmods.dev/badge/instructions/owenob1/prompt-improver/claude-md/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/instructions/owenob1/prompt-improver/claude-md"><img src="https://agentmods.dev/badge/instructions/owenob1/prompt-improver/claude-md.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 2,653 This file is loaded in full into every session.

When invoked 2,653 The same file — it is already loaded in full.

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	$0.02653	$0.02653
Opus 5.5	$0.01061	$0.01061
Sonnet 5.5	$0.00531	$0.00531
Haiku 4.5	$0.00265	$0.00265

Measured 1mo ago against content hash 938755d1c936, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

prompt-improver CLAUDE.md 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 1mo 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.

CLAUDE.md · 106 lines

How it starts

The opening of the file, as written. The whole thing — 106 lines — stays where its author put it; the contents beside it link to each section on GitHub.

CLAUDE.md

This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.

What this is

A pure-Bash Agent Skill (no build step, no package manager, no compiled code), plus mcp/, a TypeScript Cloudflare Worker that serves the skill's rules over MCP, and site/, the Astro + shadcn/ui info page that worker serves at /. It takes a vague coding request and rewrites it into a structured XML spec by shelling out to a separate coding CLI (claude, grok, codex, gemini, …) headlessly, then hands the result back to the host agent to execute.

Commands

bash tests/smoke-test.sh                          # full suite — this is what CI runs
bash skills/prompt-improver/scripts/smoke-test.sh # skill-package checks only (24 groups, offline stub CLIs)
cd mcp && npm ci && npm run typecheck && npm test  # hosted MCP server (CI job `mcp`)

There is no test-name filter. To iterate on one check, run the underlying script directly:

cd skills/prompt-improver
bash scripts/validate-prompt.sh examples/fixtures/valid-prompt.xml    # expect exit 0
bash scripts/validate-prompt.sh examples/fixtures/invalid-prompt.xml  # expect exit 1
bash scripts/assemble-generation-prompt.sh "some request"             # inspect assembled generator prompt
bash scripts/gather-context.sh .                                      # inspect deterministic context block
bash scripts/generate-prompt.sh --raw-input "..." --mode plan --skip-validate  # end-to-end, needs a CLI on PATH
bash scripts/generate-prompt.sh --mode plan --raw-input-file - <<'REQ'           # same, request on stdin (no shell expansion)
...
REQ

Group [1/8] of the skill smoke test runs bash -n over every scripts/**/*.sh, so syntax errors are caught without a separate lint step. There is no typecheck, formatter, or linter.

Architecture

Host agent session (Claude Code / Grok / …)
  └─ scripts/generate-prompt.sh          orchestrator
       ├─ lib/settings.sh                settings + model/backend resolution + limit detection
       ├─ gather-context.sh              deterministic repo facts (shell only, never AI search)
       ├─ assemble-generation-prompt.sh  references + context + raw request → one prompt file
       ├─ backends/<cli>.sh              headless invoke of the generator CLI
       └─ validate-prompt.sh             structural checks on the XML
  └─ host executes the XML (execute mode) or shows it (plan mode)

Read the full file on GitHub · 106 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.

1mo ago Changed · +16 lines · +862 tokens per session 938755d1c936
2mo ago First seen · 90 lines · 1,791 tokens per session scan A a115b35f70b3

Subscribe to this mod's changes

prompt-improver CLAUDE.md is an instructions file published in the GitHub repository owenob1/prompt-improver (6 stars, last pushed 1mo ago), licensed MIT. It adds 2,653 tokens to every session, about $0.0106 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 2026-08-31.

Building agents LLM Writing and docs agent-skills ai-agents bash cli prompt-engineering

prompt-improver CLAUDE.md next to 6 instructions from other repositories
Mod	Type	Stars	Last push	Licence	Scan	Tokens	Works with
prompt-improver CLAUDE.mdowenob1/prompt-improver	Instructions file	6	1mo ago	MIT	A	2,653	Claude Code
vscode buildNext.instructions.mdmicrosoft/vscode	Instructions file	195k	today	MIT	A	9,719	GitHub Copilot
next.js AGENTS.mdvercel/next.js	Instructions file	143k	yesterday	MIT	A	7,461	Codex, OpenCode
langchain AGENTS.mdlangchain-ai/langchain	Instructions file	149k	today	MIT	A	4,680	Codex, OpenCode
vscode oss-third-party-notices.instructions.mdmicrosoft/vscode	Instructions file	195k	today	MIT	A	5,516	GitHub Copilot
vscode copilot-instructions.mdmicrosoft/vscode	Instructions file	195k	today	MIT	A	3,093	GitHub Copilot
deepseek-harness AGENTS.mddeepseek-ai/deepseek-harness	Instructions file	275k	2d ago	MIT	A	4,313	Codex, OpenCode

prompt-improver: Instructions file for Claude Code