技能 fiduciary-duty-negligence
📦

fiduciary-duty-negligence

內容修訂 r1 高風險 ⚙️ 外部命令

分析受託義務證據

受託義務證據審查需要一致的中繼資料、引文和中立結論。此技能會引導 Claude、Codex 和 Claude Code 完成結構化的法律審查工作流程。

支援: Claude Codex Code(CC)
Maintain this Skill? Claim it ⚠️ 38 差

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 "fiduciary-duty-negligence" from https://skillstore.io/skills/acskamloops-fiduciary-duty-negligence.md and its manifest at https://skillstore.io/api/skills/acskamloops-fiduciary-duty-negligence/manifest. Verify the artifact. Stop and obtain explicit user consent before installing or changing files.

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。

Markdown 詳情 GET /skills/acskamloops-fiduciary-duty-negligence.md 已簽名 manifest GET /api/skills/acskamloops-fiduciary-duty-negligence/manifest 已簽名 lockfile GET /api/skills/acskamloops-fiduciary-duty-negligence/lockfile

測試它

正在使用「fiduciary-duty-negligence」。 一項任務包含一封關於保留地面積變更和信託帳戶入帳缺失的信件。

預期結果:

回應會識別文件中繼資料，引用面積和帳戶段落，並僅陳述事實脈絡。

正在使用「fiduciary-duty-negligence」。 一份文件有亂碼文字且缺少來源。

預期結果:

回應會將該文字標記為需要 OCR 審查，並說明來源問題，而不作出法律結論。

正在使用「fiduciary-duty-negligence」。 一份草稿在結論中使用 likely 和 suggests 等詞。

預期結果:

回應會標記意見性語言，並將結論改寫為與引文證據相連的中立陳述。

安全審計

高風險
v8 • 2026/7/6 開啟版本化報告

Most static findings are false positives caused by Markdown backticks around paths, field names, and examples. The confirmed issues are actionable local workflow commands, PowerShell command blocks, and an instruction hierarchy risk from task-provided system instructions.

1
已掃描檔案
162
分析行數
11
複核項目
0
已忽略的誤報

已確認的安全問題 (2)

高
SKILL.md:20
Prompt Injection Attempt Detected
The skill says, "You must follow the `system_instructions` block injected into every JSON task file." This delegates instruction authority to task input and could let untrusted JSON override the model workflow.
The suspicious text directly names system instructions and tells the agent to follow instructions injected into task files. This is a clear instruction hierarchy risk.
中
SKILL.md:123-128SKILL.md:130-139
Workflow Commands Mutate Files And Review State
The submit and flag workflow can append analysis to internal reports, move tasks to manager review, log flagged tasks, and move files into an OCR pipeline. These side effects are broader than read-only legal analysis.
The cited lines describe submit and flag commands plus their state-changing results. The behavior is explicit, although it may be legitimate in a trusted local workflow.
能力複核項目 (11)

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。

高
SKILL.md:35
PowerShell invocation
```powershell
The PowerShell block introduces a command that runs a local Python workflow script. It is an actionable shell-command instruction, not harmless formatting in context.
高
SKILL.md:42
PowerShell invocation
```powershell
The PowerShell block contains a python -c file-reading command with a runtime path placeholder. This is actionable shell execution in the skill workflow.
高
SKILL.md:124
PowerShell invocation
```powershell
The PowerShell block contains the submit-task command that runs local project code and appends review output. This is a state-changing shell workflow.
高
SKILL.md:133
PowerShell invocation
```powershell
The PowerShell block contains a flag-task command that executes local workflow code. It modifies queue records rather than only producing analysis text.
高
SKILL.md:138
PowerShell invocation
```powershell
The PowerShell block contains a flag-task command that can move source files into an OCR pipeline. This is a state-changing external command.
中
SKILL.md:10
Ruby/shell backtick execution
- For court audit trails, run batches via `codex_exec_runner.sh` with `PUKAIST_CODEX_LOG_EVENTS=1` t
The instruction tells the agent to run a local shell script with an environment flag for batch execution. That is real external command execution outside the skill text.
中
SKILL.md:35-37
Ruby/shell backtick execution
```powershell
This fenced workflow block contains a Python command to run refinement_workflow.py get-task. The command is operational and executes local project code.
中
SKILL.md:42-44
Ruby/shell backtick execution
```powershell
This block runs python -c to open and print an input JSON file using a runtime path. It is real external command execution, even though the purpose is data viewing.
中
SKILL.md:124-126
Ruby/shell backtick execution
```powershell
This submit-task command runs a local workflow script and writes results into project reports. It changes workflow state through external command execution.
中
SKILL.md:133-135
Ruby/shell backtick execution
```powershell
This flag-task command executes a local workflow script that logs exceptions. It changes workflow state and relies on task identifiers supplied at runtime.
中
SKILL.md:138-140
Ruby/shell backtick execution
```powershell
This flag-task command executes a local workflow script and can move files into an OCR pipeline. It has filesystem side effects beyond analysis.

風險因素

⚙️ 外部命令 (41)
SKILL.md:7 SKILL.md:8 SKILL.md:10 SKILL.md:20 SKILL.md:30 SKILL.md:31 SKILL.md:35-37 SKILL.md:37-40 SKILL.md:40-42 SKILL.md:42-44 SKILL.md:44-62 SKILL.md:62 SKILL.md:63-94 SKILL.md:94 SKILL.md:97-99 SKILL.md:99-102 SKILL.md:102-104 SKILL.md:104-109 SKILL.md:109-112 SKILL.md:112 SKILL.md:119 SKILL.md:120 SKILL.md:121-124 SKILL.md:124-126 SKILL.md:126-127 SKILL.md:127-128 SKILL.md:128 SKILL.md:132-133 SKILL.md:133-135 SKILL.md:135-137 SKILL.md:137-138 SKILL.md:138-140 SKILL.md:140-143 SKILL.md:143 SKILL.md:144-145 SKILL.md:145 SKILL.md:35 SKILL.md:42 SKILL.md:124 SKILL.md:133 SKILL.md:138
審計單位: codex 檢視審計紀錄 →
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。

開啟版本化報告
安全評估

複製報告連結

https://skillstore.io/skills/acskamloops-fiduciary-duty-negligence/audits/8?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_report

Markdown 徽章

[![Skillstore security assessment](https://skillstore.io/badges/skills/acskamloops-fiduciary-duty-negligence/security.svg)](https://skillstore.io/skills/acskamloops-fiduciary-duty-negligence?utm_source=security_passport_badge)

HTML 徽章

<a href="https://skillstore.io/skills/acskamloops-fiduciary-duty-negligence?utm_source=security_passport_badge"><img src="https://skillstore.io/badges/skills/acskamloops-fiduciary-duty-negligence/security.svg" alt="Skillstore security assessment" loading="lazy"></a>

嵌入卡片

<iframe src="https://skillstore.io/embed/skills/acskamloops-fiduciary-duty-negligence.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

ACSKamloops. (2026). fiduciary-duty-negligence security audit report (audit version 8) [Author version unspecified]. Skillstore. https://skillstore.io/skills/acskamloops-fiduciary-duty-negligence/audits/8

BibTeX 引用

@techreport{acskamloops-acskamloops-fiduciary-duty-negligence-2026, author = {ACSKamloops}, title = {fiduciary-duty-negligence security audit report (audit version 8)}, institution = {Skillstore}, year = {2026}, number = {8}, url = {https://skillstore.io/skills/acskamloops-fiduciary-duty-negligence/audits/8}, 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: "fiduciary-duty-negligence security audit report (audit version 8)" version: "unspecified" type: report authors: - name: "ACSKamloops" date-released: "2026-07-06" url: "https://skillstore.io/skills/acskamloops-fiduciary-duty-negligence/audits/8" identifiers: - type: other value: "skillstore:acskamloops-fiduciary-duty-negligence:audit:8" description: "Skillstore immutable audit report identifier"

Skillstore 評分

評分依據 證據置信度: 中
55
架構
85
可維護性
87
內容
72
社群
83
規範符合度

你可以打造什麼

索引保留地證據

使用中立證據筆記審查文件中的土地保護失敗、侵入土地問題和面積變更。

準備審查佇列項目

將任務輸入轉換為一致的中繼資料、摘要、結論和關鍵引文，以供主管審查。

標記 OCR 或來源問題

在資料進入精修證據報告之前，識別亂碼文字、缺失來源或無關材料。

試試這些提示

審查單一任務 
審查此受託義務任務。擷取文件標題、年份、來源、中立摘要和直接證據引文。
擷取含中繼資料的證據 
分析此文件文字與受託義務的相關性。包含已驗證的中繼資料、可靠性、OCR 狀態、頁面錨點和事實證據重要性。
驗證提交草稿 
檢查此草稿是否缺少中繼資料、含有意見詞、結論過短、來源薄弱，以及有無未受支持的事實主張。
稽核批次一致性 
審查此批次中的相關性判斷、中立語言、直接引文、矛盾註記，以及可供主管審閱的證據紀錄是否一致。

最佳實務

  • 起草前，請根據文件內容驗證來源、日期和標題。
  • 使用直接引文和中立的事實措辭；提交前移除意見詞。
  • 僅在使用者核准後，於受信任的工作區中執行會變更狀態的工作流程命令。

避免

  • 不要在未閱讀周邊脈絡的情況下，將關鍵字命中視為相關。
  • 不要編造日期、標題、文件識別碼或頁面錨點。
  • 不要在檢查來源和 OCR 狀態之前提交或標記任務。

常見問題

此技能是否提供法律建議？
否。它支援證據整理和中立事實起草，而非法律建議或法律策略。
它聚焦於哪些文件？
它聚焦於受託義務、保留地、信託基金、侵入土地、利益衝突和疏忽證據。
沒有本機工作流程腳本也能運作嗎？
它可以引導分析，但 fetch、submit 和 flag 命令取決於本機專案腳本。
為什麼它要求中立措辭？
此工作流程是為事實證據紀錄而設計，因此結論必須避免未受支持的意見。
應如何處理不確定的中繼資料？
審查者應檢查頁首、頁尾、檔案名稱和內容，然後才將文件標記為未註明日期。
這對自動執行是否安全？
否。它包含本機命令，只有在受控工作區中取得明確核准後才應執行。

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。

儲存庫

https://github.com/ACSKamloops/shs-engine/tree/master/.codex/skills/fiduciary-duty-negligence

引用

c1fdca50ff516318f65fed0d7f9e82797c5171dc

維護新鮮度

2026/7/18

使用量

12 次下載 · 329 次瀏覽

檔案結構

📄 SKILL.md