📦
planning-with-files
83使用規劃檔案組織長時間 AI 任務
由 ZhanlinCui
長時間的 AI 工作階段可能會遺失目標、決策和進度。此技能會引導 Claude、Codex 和 Claude Code 將計畫、發現和記錄儲存在專案檔案中。
效率 查看
技能 syncbook
📦
syncbook
內容修訂 r1 中風險 🌐 網路存取📁 檔案系統存取⚙️ 外部命令🔑 環境變數
Build Reciprocal Connections with Syncbook
Finding collaborators requires clear offers, needs, and current project context. Syncbook helps your agent prepare reviewed profiles, explore reciprocal connections, and organize authorized follow-up.
支援: Claude Codex Code(CC)
使用我的智慧代理安裝
將此請求複製給你的智慧代理，其中包含權威 Skill 頁面和清單。
智能體請求
Review the Skillstore skill "syncbook" from https://skillstore.io/skills/theguysaccount-syncbook.md and its manifest at https://skillstore.io/api/skills/theguysaccount-syncbook/manifest. Verify the artifact. You may proceed after verification, subject to the environment's own policy.你的智慧代理仍應顯示其計畫，並依安全政策請求必要確認。
Agent 可讀資源
當 AI Agent、爬蟲或腳本需要乾淨脈絡、而不是讀取完整頁面時，請使用這些連結。
測試它
正在使用「syncbook」。 Prepare a profile from my community garden project brief without submitting anything.
預期結果:
- Prepared locally; nothing submitted or published.
- Direction: organize a neighborhood garden pilot.
- Offer: volunteer coordination and planting plans.
- Need: an approved growing space and a partner for a small pilot.
- Review needed: confirm available time and which details may be public.
正在使用「syncbook」。 Run my weekly review when no scheduler or project source is available.
預期結果:
- No scheduled task was created.
- Progress could not be verified because the authorized project source is unavailable.
- A manual review can continue after you provide the approved source.
- No introductions or external messages were sent.
正在使用「syncbook」。 Report this invitation failure, but feedback sharing is disabled.
預期結果:
- A short capability-gap summary is prepared locally.
- Nothing was submitted because feedback sharing is disabled.
- You can enable sharing in your workspace or keep the summary local.
安全審計
中風險v1 • 2026/10/25 開啟版本化報告
Most static detections are benign documentation, JavaScript syntax, or private credential storage; no evidence of deliberate credential theft was found. The shared request helper can forward sensitive bodies through redirects, and remote prompts introduce an unpinned instruction trust boundary.
6
已掃描檔案
308
分析行數
1
複核項目
0
已忽略的誤報
已確認的安全問題 (1)
中
Unpinned Remote Instructions Enter Agent Workflows
The skill says to read the 'complete coordination handoff' and create a weekly job using a live remote prompt. No pinned version or review step protects against changed instructions entering a persistent task with authorized connectors.
The source explicitly delegates coordination and scheduled instructions to mutable remote text without integrity checks. No evidence of an actual malicious remote payload was found.
能力複核項目 (1)
這些是真實的本機能力，對此技能而言可能是預期行為，因此需要複核，但不會按已確認的惡意行為計分。
中
Fetch API call
async function call(path,method='GET',data,privateAccess=false){const headers={};if(data!==undefined
Fetch follows redirects without rechecking the origin allowlist. A 307 or 308 response can forward profile data or the status request's token body externally.
風險因素
🌐 網路存取 (12)
⚙️ 外部命令 (14)
審計單位: codex
分享與引用此報告
分享版本化評估報告、中性徽章、嵌入卡片與引用。Skillstore 只報告證據，不判斷該 Skill 是否安全。
複製報告連結
https://skillstore.io/skills/theguysaccount-syncbook/audits/1?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_reportMarkdown 徽章
[](https://skillstore.io/skills/theguysaccount-syncbook?utm_source=security_passport_badge)HTML 徽章
<a href="https://skillstore.io/skills/theguysaccount-syncbook?utm_source=security_passport_badge"><img src="https://skillstore.io/badges/skills/theguysaccount-syncbook/security.svg" alt="Skillstore security assessment" loading="lazy"></a>嵌入卡片
<iframe src="https://skillstore.io/embed/skills/theguysaccount-syncbook.html" title="Skillstore Security Assessment" sandbox="allow-popups allow-popups-to-escape-sandbox" loading="lazy" referrerpolicy="no-referrer" width="420" height="180"></iframe>學術引用 (APA · BibTeX · CFF)
APA 引用
theguysaccount. (2026). syncbook security audit report (audit version 1) [Author version unspecified]. Skillstore. https://skillstore.io/skills/theguysaccount-syncbook/audits/1BibTeX 引用
@techreport{theguysaccount-theguysaccount-syncbook-2026,
author = {theguysaccount},
title = {syncbook security audit report (audit version 1)},
institution = {Skillstore},
year = {2026},
number = {1},
url = {https://skillstore.io/skills/theguysaccount-syncbook/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: "syncbook security audit report (audit version 1)"
version: "unspecified"
type: report
authors:
- name: "theguysaccount"
date-released: "2026-10-25"
url: "https://skillstore.io/skills/theguysaccount-syncbook/audits/1"
identifiers:
- type: other
value: "skillstore:theguysaccount-syncbook:audit:1"
description: "Skillstore immutable audit report identifier"
Skillstore 評分
評分依據 證據置信度: 中 77
架構
85
可維護性
87
內容
65
社群
91
規範符合度
你可以打造什麼
Prepare a Founder Profile
Turn an authorized project brief into clear offers, needs, and collaboration goals, then review the profile before approval.
Review Weekly Collaboration Options
Review actual project progress and up to three reciprocal fits using a separately authorized weekly routine.
Report Product Friction
Prepare a concise capability-gap report, verify feedback access and opt-in, and submit only the approved summary.
試試這些提示
Prepare My First Profile
Help me prepare a Syncbook profile using only this project brief. Show the draft before submitting it and keep unrelated context out.
Compare Reciprocal Connections
Using my approved Syncbook access, compare up to three reciprocal matches. Separate public evidence from assumptions and suggest a bounded first step.
Prepare Authorized Feedback
Prepare a short Syncbook capability-gap report about this failed action. Show it first, verify feedback scope and opt-in, and exclude private conversation details.
Configure a Weekly Review
Prepare a Friday 09:00 America/Los_Angeles review using only my approved project source. Show the fetched weekly prompt for approval before scheduling. Verify the task receipt.
最佳實務
- Provide a narrow authorized source and review profile content before approving any publication.
- Use short-lived, purpose-specific tokens and keep them in private state or the host credential store.
- Review remote prompts before scheduling, verify task and submission receipts, and request redirect hardening before transmitting sensitive data.
避免
- Include raw conversations, unrelated private facts, or other people's contact details in profile or feedback payloads.
- Place tokens in prompts, command arguments, repositories, or public review links.
- Treat copied prompts, share sheets, or attempted requests as proof of scheduling, delivery, or successful submission.
常見問題
Which agents can use this skill?
The skill targets Claude, Codex, and Claude Code. Available connectors, terminal access, scheduling, and delivery tools vary by host.
Does installation create a Syncbook account?
No. Installation provides instructions and a terminal helper. Signup still requires a private review flow and human approval.
What does the terminal helper require?
The dependency-free helper requires Node 20 or later. It reads explicitly supplied files and its own private Syncbook state.
How are private operations authorized?
They require owner-approved scoped tokens. The helper accepts existing tokens through stdin and stores them with private file permissions.
Can it run weekly reviews automatically?
Only with an authorized native scheduler, agreed sources, active preferences, and appropriate access. Without a scheduler, it offers manual reviews.
What security concerns did the audit identify?
Automatic redirects can forward sensitive API bodies outside the allowlist. Live coordination and weekly prompts lack pinned versions or mandatory review.
開發者詳細資訊
授權
MIT
Skillstore 修訂
r1
版本提示
作者未聲明版本。
引用
83d8a0a42a32126fb19983512e32b45bfe3d70f9
維護新鮮度
2026/10/25
使用量
0 次下載 · 0 次瀏覽
相關技能
查看全部📦
context-optimization
80最佳化 AI 上下文視窗
由 ChakshuGautam
長時間 AI 工作階段可能浪費詞元、拖慢回應速度，並遺失重要上下文。這項技能為 Claude、Codex 和 Claude Code 提供實用模式，用於壓縮、遮罩、快取和規劃上下文預算。
效率 查看
📦
verification-protocol
80以證據驗證代理完成狀態
由 CleanExpo
AI 代理可能在沒有證明的情況下將工作標記為完成，造成虛假完成的風險。此技能要求在接受任務完成前提供獨立證據。
效率 查看
📦
brainstorming
79將想法塑造成經驗證的設計
由 ZhanlinCui
早期產品與工程想法經常失敗，是因為需求一直模糊不清。此技能會引導 Claude、Codex 或 Claude Code，在實作前透過聚焦問題、選項與經驗證的設計來推進。
效率 查看
📦
bv
79使用圖形指標排定 Beads 工作優先順序
由 Dicklesworthstone
團隊手動閱讀 issue 清單時，常會漏掉阻塞項目與循環。此 skill 會將 Beads 資料轉化為聚焦的分流、規劃與 agent 工作流程指引。
效率 查看
📦
gauntlet
90Review Defects and Verify Root-Cause Repairs
由 aakashh242
Reviews can miss failure conditions or accept fixes without reliable evidence. Gauntlet guides scoped adversarial reviews, authorized repairs, and revision-bound verification within a finite budget.
程式開發 查看