Skip to main content

build-a-data-retention-schedule

Build an operational schedule for retaining and disposing of records. Use when systems hold overlapping record types with legal, contractual, business, privacy, preservation, and deletion constraints.

来源信息

仓库
protosphinx/sphinxstack
最近来源活动
2026年7月29日 13:55
检测到的 SKILL.md 语言
英语
星标
3
Forks
0

安装方式

默认使用会先检查来源的 Prompt;你也可以切换为直接命令,或下载本地副本。

检查来源文件

决定是否安装前,请先阅读 SKILL.md,以及 SkillsMP 当前展示的配套文件。

文件资源管理器
2 个文件

正在显示 SKILL.md

SKILL.md
来源说明 · 只读预览
name
build-a-data-retention-schedule
category
data
description
Build an operational schedule for retaining and disposing of records. Use when systems hold overlapping record types with legal, contractual, business, privacy, preservation, and deletion constraints.
# build-a-data-retention-schedule Create rules that systems can execute and owners can defend. Avoid one blanket period for a database containing different record classes. ## Inputs - Record inventory, purposes, systems, owners, and data flows - Applicable requirements supplied by authorized legal, privacy, security, finance, or records owners - Contract terms, investigation holds, recovery dependencies, and deletion capabilities - Current retention settings, backups, archives, replicas, exports, and vendor behavior ## Procedure 1. Define scope, authoritative policy sources, approval roles, and the date on which requirements were checked. 2. Group records by business event and purpose, not just table or file name. 3. Record the retention trigger, minimum, maximum, rationale, authority, and disposition for each class. 4. Separate active use, inactive archive, legal hold, backup expiry, anonymization, and final deletion. 5. Identify conflicts and unknowns without choosing a legal interpretation on your own. 6. Map each record class to every system copy, processor, export, cache, index, and recovery store. 7. Specify machine-testable start events, deletion jobs, exceptions, approvals, evidence, and failure alerts. 8. Test representative records across clock boundaries, linked objects, holds, restore paths, and vendor deletion. 9. Reconcile actual age distributions and deletion logs against the schedule. 10. Approve, version, publish, and review the schedule when purpose, system, or authority changes. ## Guardrails - Never delete records subject to a confirmed hold or preservation duty. - Do not retain everything “just in case”; record the purpose and authority. - Do not claim hard deletion when backups, processors, or derived stores remain. - Use privacy-minimized samples when testing. ## Done - Every in-scope record class has a trigger, period, authority, owner, and disposition - System copies and exceptions are mapped - Deletion and hold behavior has been tested - Reconciliation evidence and review triggers are retained
在 GitHub 查看