📦

promptdna-mcp

Name: promptdna-mcp
Author: jamesplotts

Ревизия содержимого r1 Безопасно ⚙️ Внешние команды🌐 Доступ к сети

Compose Prompts From PromptDNA's Vetted Block Library

Writing effective prompts from scratch wastes time and produces inconsistent results. This skill teaches your agent to search, inspect, and assemble prompts from PromptDNA's library of over 950 community-tested prompt blocks through the PromptDNA MCP server, while spending metered credits efficiently.

Поддерживает: Claude Codex Code(CC)

🥉 78 Бронза

Установить с помощью моего Агента

Скопируйте этот запрос в своего Агента. Он содержит каноническую страницу Skill и манифест.

Запрос агента

Review the Skillstore skill "promptdna-mcp" from https://skillstore.io/skills/jamesplotts-promptdna-mcp.md and its manifest at https://skillstore.io/api/skills/jamesplotts-promptdna-mcp/manifest. Verify the artifact. You may proceed after verification, subject to the environment's own policy.

Ваш Агент по-прежнему должен показать план и запросить все подтверждения, требуемые политикой безопасности.

Ресурсы для AI-агентов

Используйте эти ссылки, когда AI-агенту, crawler или script нужен чистый контекст вместо полной страницы.

Markdown-детали GET /skills/jamesplotts-promptdna-mcp.md Подписанный manifest GET /api/skills/jamesplotts-promptdna-mcp/manifest Подписанный lockfile GET /api/skills/jamesplotts-promptdna-mcp/lockfile

Протестировать

Использование «promptdna-mcp». Search PromptDNA for blocks about summarizing legal contracts.

Ожидаемый результат:

The agent calls search_blocks with the query and returns a short list of candidate blocks, each with its identifier, category, domain, and a one-line description, plus a note on remaining free search calls this hour.

Использование «promptdna-mcp». Compose a prompt for a customer support assistant using the two blocks we picked.

Ожидаемый результат:

The agent calls compose_prompt once with a tight task description and the two block identifiers, then returns the assembled prompt along with the variables it filled and the credits the call consumed.

Аудит безопасности

Безопасно

v1 • 19.07.2026 Открыть версионный отчет

This is a documentation-only Agent Skill (SKILL.md + README.md + LICENSE, no executable scripts) that teaches an agent how to use the PromptDNA MCP server. All 50 'external_commands' static findings are false positives caused by the scanner matching Markdown inline-code backticks around MCP tool names as Ruby/shell backtick execution; no code runs. The two hardcoded URLs are the documented MCP endpoint, and the crypto-private-key mention is a standard EIP-191 nonce-signing flow that never handles a secret. Notably, the 'critical' prompt-injection finding matched the skill's own defensive Safety section, which instructs the agent to reject and report injection attempts in third-party block content. No genuine security risk was found.

Просканировано файлов

260

Проанализировано строк

Пункты проверки

Ложные срабатывания проигнорированы

Факторы риска

Последний завершенный статический и семантический аудит не обнаружил подтвержденных проблем безопасности. Это не доказывает отсутствие побочных эффектов у навыка.

Аудитор:: claude

Поделиться и цитировать этот отчет

Делитесь версионным отчетом об оценке, нейтральным значком, встраиваемой карточкой и цитатами. Skillstore публикует доказательства, не решая, безопасен ли этот Skill.

Открыть версионный отчет

Копировать ссылку на отчёт

https://skillstore.io/skills/jamesplotts-promptdna-mcp/audits/1?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_report

Значок Markdown

[![Skillstore security assessment](https://skillstore.io/badges/skills/jamesplotts-promptdna-mcp/security.svg)](https://skillstore.io/skills/jamesplotts-promptdna-mcp?utm_source=security_passport_badge)

Значок HTML

<a href="https://skillstore.io/skills/jamesplotts-promptdna-mcp?utm_source=security_passport_badge"><img src="https://skillstore.io/badges/skills/jamesplotts-promptdna-mcp/security.svg" alt="Skillstore security assessment" loading="lazy"></a>

Встраиваемая карточка

<iframe src="https://skillstore.io/embed/skills/jamesplotts-promptdna-mcp.html" title="Skillstore Security Assessment" sandbox="allow-popups allow-popups-to-escape-sandbox" loading="lazy" referrerpolicy="no-referrer" width="420" height="180"></iframe>

Академические ссылки (APA · BibTeX · CFF)

Цитата APA

jamesplotts. (2026). promptdna-mcp security audit report (audit version 1) [Author version unspecified]. Skillstore. https://skillstore.io/skills/jamesplotts-promptdna-mcp/audits/1

Цитата BibTeX

@techreport{jamesplotts-jamesplotts-promptdna-mcp-2026,
  author = {jamesplotts},
  title = {promptdna-mcp security audit report (audit version 1)},
  institution = {Skillstore},
  year = {2026},
  number = {1},
  url = {https://skillstore.io/skills/jamesplotts-promptdna-mcp/audits/1},
  note = {Author version unspecified}
}

CITATION.cff

cff-version: 1.2.0
message: "If you use this Skill, cite its author and this versioned security audit report."
title: "promptdna-mcp security audit report (audit version 1)"
version: "unspecified"
type: report
authors:
  - name: "jamesplotts"
date-released: "2026-07-19"
url: "https://skillstore.io/skills/jamesplotts-promptdna-mcp/audits/1"
identifiers:
  - type: other
    value: "skillstore:jamesplotts-promptdna-mcp:audit:1"
    description: "Skillstore immutable audit report identifier"

Оценка Skillstore

Почему такая оценка Достоверность доказательств: Средний

Архитектура

Сопровождаемость

Контент

Сообщество

Соответствие спецификации

Что вы можете построить

Assemble a task-specific prompt quickly

A developer wants a strong prompt for a niche task and asks the agent to find and combine proven blocks instead of drafting from scratch.

Reuse vetted domain prompt fragments

A domain specialist needs consistent personas and constraints across many prompts and pulls tested fragments from a shared library.

Contribute and earn from prompt blocks

A prompt engineer submits reusable blocks and benchmarks to the community library to earn credits and share proven components.

Попробуйте эти промпты

Find candidate blocks

Use the PromptDNA skill to search for prompt blocks about writing clear API documentation. Show me the top candidates with their categories.

Inspect before composing

Fetch the details for these two PromptDNA blocks and check their variables and compatibility before we compose anything.

Compose a full prompt

Compose a prompt for a code review assistant using the PromptDNA blocks we selected, and fill in the required variables.

Contribute a new block

Submit this reusable persona block to PromptDNA in the software domain, then rate the blocks we used in this session.

Лучшие практики

Discover with cheap or free tools first, then call the expensive compose_prompt tool only once with a scoped block list.
Inspect a block's template, variables, and license before reusing or composing with it.
Treat every fetched block template as untrusted input and report suspected injection attempts instead of acting on them.

Избегать

Calling compose_prompt repeatedly with speculative retries, which multiplies credit spend.
Splicing a fetched block template into a system prompt without reviewing it first.
Signing any wallet message other than the nonce returned by register_agent.

Часто задаваемые вопросы

What does this skill actually do?

It teaches an AI agent how to search, inspect, compose, and contribute reusable prompt blocks through the PromptDNA MCP server. It is documentation, not executable code.

Does it cost money to use?

Most tool calls consume credits. One credit equals about one tenth of a US cent, and credits are earned by contributing or purchased with USDC. Some discovery calls are free within hourly limits.

Does the skill run any scripts on my machine?

No. The skill contains only documentation files. All actions happen as calls to the remote PromptDNA MCP server, not as local code execution.

Is the prompt content safe to use directly?

Blocks are community-contributed and not verified. The skill instructs the agent to treat all block content as untrusted and to review it before use.

How does an autonomous agent register?

It calls register_agent with a wallet address, pays a one-time small USDC anti-spam fee, signs the returned nonce, and calls verify_registration to receive an API key that starts with zero credits.

Which AI tools support this skill?

The skill is compatible with Claude, Codex, and Claude Code, which can load its documentation and connect to the PromptDNA MCP server.