skill-harvest
Discover Reusable Skills in Your Session History
Useful methods and overlooked installed skills are scattered across past conversations. Skill Harvest organizes Claude Code history into evidence for skill discovery, testing, and approved publication.
Установить с помощью моего Агента
Скопируйте этот запрос в своего Агента. Он содержит каноническую страницу Skill и манифест.
Review the Skillstore skill "skill-harvest" from https://skillstore.io/skills/rabbitdesigns22-skill-harvest.md and its manifest at https://skillstore.io/api/skills/rabbitdesigns22-skill-harvest/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.
Протестировать
Использование «skill-harvest». Illustrative request: find reusable methods from recent sessions.
Ожидаемый результат:
- Candidate: deployment verification.
- Evidence: three independent requests across two sessions, including repeated corrections about checking the live result.
- Next step: compare public alternatives before deciding whether to build.
Использование «skill-harvest». Illustrative request: check whether installed skills were overlooked.
Ожидаемый результат:
- Possible missed skill: diagnosing-bugs.
- Reason: repeated debugging requests matched its trigger phrases, but the session recorded no load of that skill.
- Action: review relevance and skill availability at the time before sending a hint.
Использование «skill-harvest». Illustrative request: publish the strongest candidate.
Ожидаемый результат:
- Publication paused: the candidate has evidence from only one session on one day.
- It does not meet the independent-evidence gate.
- No repository was published.
Аудит безопасности
Средний рискMost static findings identify harmless Markdown formatting, documented local skill paths, or public GitHub inventory requests. Confirmed output risks concern unredacted private digests and spreadsheet formulas in CSV metadata; a separate finding covers unpinned npm execution. No evidence found of malicious prompt injection or covert exfiltration; publication and hook installation explicitly require user approval.
Подтверждённые проблемы безопасности (1)
Пункты проверки возможностей (2)
Это реальные локальные возможности, которые могут ожидаться для этого навыка, поэтому они требуют проверки, но не считаются подтверждённым вредоносным поведением.
Факторы риска
⚙️ Внешние команды (21)
📁 Доступ к файловой системе (16)
🌐 Доступ к сети (1)
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.
Копировать ссылку на отчёт
https://skillstore.io/skills/rabbitdesigns22-skill-harvest/audits/1?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_reportЗначок Markdown
[](https://skillstore.io/skills/rabbitdesigns22-skill-harvest?utm_source=security_passport_badge)Значок HTML
<a href="https://skillstore.io/skills/rabbitdesigns22-skill-harvest?utm_source=security_passport_badge"><img src="https://skillstore.io/badges/skills/rabbitdesigns22-skill-harvest/security.svg" alt="Skillstore security assessment" loading="lazy"></a>Встраиваемая карточка
<iframe src="https://skillstore.io/embed/skills/rabbitdesigns22-skill-harvest.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
rabbitdesigns22. (2026). skill-harvest security audit report (audit version 1) [Author version unspecified]. Skillstore. https://skillstore.io/skills/rabbitdesigns22-skill-harvest/audits/1Цитата BibTeX
@techreport{rabbitdesigns22-rabbitdesigns22-skill-harvest-2026,
author = {rabbitdesigns22},
title = {skill-harvest security audit report (audit version 1)},
institution = {Skillstore},
year = {2026},
number = {1},
url = {https://skillstore.io/skills/rabbitdesigns22-skill-harvest/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: "skill-harvest security audit report (audit version 1)"
version: "unspecified"
type: report
authors:
- name: "rabbitdesigns22"
date-released: "2026-11-11"
url: "https://skillstore.io/skills/rabbitdesigns22-skill-harvest/audits/1"
identifiers:
- type: other
value: "skillstore:rabbitdesigns22-skill-harvest:audit:1"
description: "Skillstore immutable audit report identifier"
Оценка Skillstore
Почему такая оценка Достоверность доказательств: СреднийЧто вы можете построить
Find Repeated Development Methods
Review recurring requests and corrections across sessions to identify reusable skills supported by independent evidence.
Improve Installed Skill Usage
Find installed skills that were not loaded during relevant work, then review targeted reminders before enabling them.
Evaluate Skill Publication Candidates
Compare candidate methods with public alternatives, test three past prompts, and prepare sanitized materials for explicit approval.
Попробуйте эти промпты
Review my last seven days of Claude Code sessions for repeated methods. Keep results private and do not install or publish anything.
Compare my recent session digests with installed skills. Show likely missed matches with dates and evidence, and manually reject irrelevant matches.
Evaluate the strongest candidate using independent session evidence and public alternatives. Use generic search terms without private data and request approval before package execution.
Draft the qualifying skill, compare three sanitized prompts with and without it, and propose reminder rules. Wait for approval before installation or publication.
Лучшие практики
- Use a private output directory, review digests for secrets, and neutralize formula-like CSV cells before spreadsheet use.
- Count independent evidence, verify heuristic matches manually, and compare public alternatives using generic search terms.
- Approve package execution, hook settings, and each publication separately after reviewing the proposed changes.
Избегать
- Treating repeated copies of one broadcast message as independent evidence for a new skill.
- Publishing transcript excerpts or assuming generated digests have already removed confidential information.
- Installing every matched skill or enabling reminder hooks without reviewing relevance and settings changes.
Часто задаваемые вопросы
Does it support Codex session history?
Does it watch every conversation live?
Does it change the original transcripts?
Are private details removed automatically?
Will it publish or install hooks automatically?
What qualifies a candidate for building?
Сведения для разработчиков
Автор
rabbitdesigns22Лицензия
Неизвестно
Ревизия Skillstore
r1
Примечание о версии
Автор не указал версию.
Ссылка
f38ac41a6255e57978dd6f276aec539d5e9b5255
Актуальность поддержки
12.11.2026
Использование
0 загрузок · 0 просмотров
Структура файлов