Context Mode 上下文清理工具

Model: qwen-max | ¥0.15/call

工程方法论GPT-4.1工程实践上下文清理工具

上下文清理工具：mksglu/context-mode: ctx-purge，适用于工程实践、代码质量与开发流程优化。

Calls: 1

Skill Documentation

Context Mode 上下文清理工具

摘要

上下文清理工具：mksglu/context-mode: ctx-purge，适用于工程实践、代码质量与开发流程优化。

> 来源: mksglu/context-mode (19.7k stars) — Claude Code 上下文窗口优化

> 原文件: skills/ctx-purge/SKILL.md

> 模型推荐: gpt-4.1 (工具/查询类)

这个 skill 是干嘛的

mksglu 在 19.7k stars 的 context-mode 仓库里整理的"上下文窗口优化"工具集 — 通过 sandbox 输出压缩(98% reduction)、会话记忆持久化、17 平台路由,让 Claude 在长对话中保持高效。

michael 强调"skill 要有相应的指导功能,指导用户使用",所以加了下面两节让 Agent 和用户对接。

---

🤖 Agent 使用说明

1. 用户提到"上下文 / context / 长会话 / 记忆"时触发

2. skill 会按"压缩 → 持久化 → 检索"流程跑

3. 涉及外部存储前 ask user 确认

4. 完工后让 Agent 跑自检

👤 用户需要做什么?

1. 告诉 Agent 你要管理什么对话(长会话 / 项目 / 知识库)

2. 如果需要外部存储 / API key,按提示提供

3. 全程 Agent 自动化,你只需回答授权类问题

---

原 skill 内容(mksglu/context-mode/skills/ctx-purge/SKILL.md,截断到 12k chars)

---

name: ctx-purge

description: |

Purge the context-mode knowledge base. Permanently deletes all indexed content

and resets session stats. This is destructive and cannot be undone.

Trigger: /context-mode:ctx-purge

user-invocable: true

---

Context Mode Purge

Permanently deletes session data for this project. Two scopes are supported (issue #520):

**Project scope** (`scope: "project"`): wipes EVERYTHING — knowledge base, all session DB rows for every session, events markdown, and stats.
**Session scope** (`sessionId: "<id>"` or `scope: "session"`): wipes ONLY the matching session's rows + FTS5 chunks. Sibling sessions, project stats, and the FTS5 store file are preserved.

Instructions

1. **Decide the scope first** with the user:

"Wipe just one session?" → ask for the `sessionId`.
"Wipe the whole project?" → confirm scope:'project' (this is the destructive, irreversible default).

2. **Warn the user about scope:'project'**. Everything will be deleted:

FTS5 knowledge base (all indexed content from `ctx_index`, `ctx_fetch_and_index`, `ctx_batch_execute`)
Session events DB (analytics, metadata, resume snapshots) for ALL sessions in the project
Session events markdown file
In-memory session stats + persisted stats file

3. Call the `mcp__context-mode__ctx_purge` MCP tool with the chosen parameters:

Scoped: `{ confirm: true, sessionId: "<id>" }` — implies scope:'session'.
Project: `{ confirm: true, scope: "project" }` — explicit destructive form.
Bare `{ confirm: true }` still works but emits a deprecation warning. Prefer the explicit forms.

4. Report the result to the user — the response lists exactly what was deleted and (for scoped purges) confirms that other sessions and project stats were preserved.

Schema rules

`confirm: true` is always required.
`sessionId` and `scope: "project"` together is REJECTED as ambiguous (the sessionId implies session scope; combining with project scope contradicts intent).
`scope: "session"` without `sessionId` throws — sessionId is required.

When to Use

**Scoped (per-session)**: scratch acceptance scenarios, drill replays, isolating a polluted session without losing the main working session's stats.
**Project**: KB contains stale or incorrect content polluting search results, switching between unrelated projects in the same session, completely fresh start.

Important

`ctx_purge` is the **only** way to delete session data. No other mechanism exists.
`ctx_stats` is read-only — shows statistics only.
`/clear` and `/compact` do NOT affect any context-mode data.
There is no undo. Re-index content if you need it again.

常见问题（FAQ）

使用「上下文清理工具」这个 skill 能解决什么问题?

本 skill 专注于上下文清理工具，mksglu/context-mode: ctx-purge。它将相关流程标准化，帮助用户更快拿到可靠结果，减少重复手工操作。

什么情况下适合使用「上下文清理工具」?

当你需要在上下文清理工具相关工作中获得稳定、可复用的产出时最适合——无论是单次任务还是纳入日常工作流，都能直接调用。

使用「上下文清理工具」前需要准备什么?

需要一个具体的项目或任务上下文，最好带有代码仓库或需求文档。

FAQ