ogham-maintain

分类: 数据与AI | 上传者: ogham-mcp | 下载: 0 | 版本: v1.0（最新）

Ogham 共享内存的管理和维护工作流程。当用户想要清理记忆、查看知识图、检查记忆统计数据、导出大脑、在切换提供商后重新嵌入记忆或回填链接时使用。触发“清理我的记忆”、“记忆统计”、“多少记忆”、“导出我的大脑”、“导出记忆”、“查看知识图”、“重新嵌入”、“链接取消链接”、“回填链接”、“记忆健康”、“奥格姆统计”、“清理过期”、“压缩旧记忆”、“压缩记忆”或任何针对奥格姆的管理/维护请求。需要连接 Ogham MCP 服务器。 --- # Ogham 维护您处理 Ogham 共享内存的管理任务。其中大部分都是不频繁的操作——提供商切换、批量清理、运行状况检查。 ## 可用操作 ### 健康检查如果用户报告问题，请先运行“health_check”。它测试数据库连接、嵌入提供程序和配置。清楚地报告发现的内容 - 如果有问题，请说明原因并提出修复建议。 ### 统计概述运行 `get_stats` 和 `list_profiles` 向用户提供其内存的图片： - 总内存和按配置文件细分 - 热门来源（客户端正在存储的） - 热门标签（哪些类别占主导地位） - 如果用户询问性能，则通过 `get_cache_stats` 缓存统计数据将其呈现为简洁的摘要，而不是原始 JSON。 ### 清理过期内存 1. 运行 `get_stats` 以显示存在多少内存 2. 检查是否有任何配置文件设置了 TTL（此信息来自 `list_profiles`） 3. 如果有过期内存，请在运行 `cleanup_expired` 之前告诉用户有多少内存 4. 仅在与用户确认后才运行“cleanup_expired”——删除是永久性的 ### Export 以用户想要的格式（JSON 或 Markdown）运行“export_profile”。告诉他们输出的去向以及如何使用它。如果他们想要导出特定的配置文件，请先使用“switch_profile”切换到该配置文件，导出，然后切换回来。 ### 重新嵌入所有记忆切换嵌入后需要这个

更新日志: Source: GitHub https://github.com/ogham-mcp/ogham-mcp

目录结构

当前层级: tree/main/skills/ogham-maintain/

📄 SKILL.md 4.2 KB

SKILL.md

---
name: ogham-maintain
description: |
  Admin and maintenance workflows for Ogham shared memory. Use when the user wants
  to clean up memories, review their knowledge graph, check memory stats, export
  their brain, re-embed memories after switching providers, or backfill links.
  Triggers on "clean up my memory", "memory stats", "how many memories",
  "export my brain", "export memories", "review knowledge graph", "re-embed",
  "link unlinked", "backfill links", "memory health", "ogham stats",
  "cleanup expired", "condense old memories", "compress memories",
  or any admin/maintenance request for Ogham.
  Requires the Ogham MCP server to be connected.
---

# Ogham maintenance

You handle admin tasks for Ogham shared memory. Most of these are infrequent operations -- provider switches, bulk cleanup, health checks.

## Available operations

### Health check

Run `health_check` first if the user reports problems. It tests database connectivity, embedding provider, and configuration. Report what it finds plainly -- if something is broken, say what and suggest a fix.

### Stats overview

Run `get_stats` and `list_profiles` to give the user a picture of their memory:
- Total memories and breakdown by profile
- Top sources (which clients are storing)
- Top tags (what categories dominate)
- Cache stats via `get_cache_stats` if they ask about performance

Present it as a concise summary, not raw JSON.

### Cleanup expired memories

1. Run `get_stats` to show how many memories exist
2. Check if any profiles have TTLs set (this info comes from `list_profiles`)
3. If there are expired memories, tell the user how many before running `cleanup_expired`
4. Run `cleanup_expired` only after confirming with the user -- deletion is permanent

### Export

Run `export_profile` with the format the user wants (JSON or Markdown). Tell them where the output goes and how to use it.

If they want to export a specific profile, switch to it first with `switch_profile`, export, then switch back.

### Re-embed all memories

This is needed after switching embedding providers (e.g. Ollama to OpenAI). It regenerates every vector in the active profile.

Before running:
1. Confirm the user has switched providers in their config
2. Warn that this takes time -- roughly 100ms per memory with a remote provider
3. Run `re_embed_all` which reports progress as it goes

After: suggest running `link_unlinked` to rebuild the knowledge graph with the new embeddings, since similarity scores will be different.

### Backfill knowledge graph links

`link_unlinked` scans memories that don't have relationship edges yet and creates links where embedding similarity is above threshold.

- Default threshold 0.85 is conservative -- only very similar memories get linked
- Suggest 0.7 for broader connections in diverse collections
- The user can set `batch_size` to control how many are processed at once

Report how many links were created when it finishes.

### Condense old memories

`compress_old_memories` shrinks old, inactive memories to save space and reduce search noise.

Three levels:
- Full text (default, recent memories)
- Condensed (key sentences, code blocks preserved, ~30% of original)
- Trace (one-line summary with tags)

Before running:
1. Explain that condensing is based on age and activity -- important, frequently-accessed, or high-confidence memories resist condensing
2. Explain that original content is always preserved and can be restored
3. Run `compress_old_memories` -- it reports how many were condensed at each level

### Profile management

- `list_profiles` -- show all profiles with memory counts
- `switch_profile` -- change active profile (session only)
- `set_profile_ttl` -- set auto-expiry. Explain that expired memories are filtered from searches immediately but not deleted until `cleanup_expired` runs
- To remove a TTL, call `set_profile_ttl` with `ttl_days=None`

## General approach

These are power-user operations. Be direct about what each one does, what it costs (time, data loss), and whether it's reversible. Deletion and re-embedding are not reversible. Exports, stats, and health checks are read-only and safe to run anytime.

If the user asks for something vague like "clean up my ogham", start with stats to understand what they have, then suggest specific actions based on what you see.

登录后下载/点赞/收藏 ❤ 83 | ★ 0

ogham-maintain

目录结构

SKILL.md

举报内容

提示