pmm-dump

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

在终端中将 PMM 内存状态转储为 ASCII art。三个深度级别：状态（仅限热图）、摘要（热图 + 集群 + 时间线）、详细（完整 ASCII）。作为子代理运行。当用户运行 /pmm-dump 或要求基于文本的内存概览时使用。触发：“pmm-dump”、“/pmm-dump”、“转储内存”、“ascii 内存”、“文本内存概述”、“显示内存热图”、“内存转储”或任何基于文本的 ASCII 内存状态可视化请求。 --- # PMM Dump 将 PMM 内存状态渲染为内联 ASCII 可视化。作为子代理运行以保持主上下文干净。 **深度级别：** $ARGUMENTS（默认：`status`，如果为空或未提供） ## 调用 - `/pmm-dump` 或 `/pmm-dump status` — 仅热图（状态级别） - `/pmm-dump 摘要` — 热图 + 集群列表 + 最后 5 个时间线条目 - `/pmm-dumpDetailed` — 完整 ASCII：图表 + 热图 + 相似度矩阵 + 集群 ## 行为使用“内存/config.md”中的“只读代理模型”（默认：“haiku”），提示如下。将“<level>”替换为深度级别（“status”、“summary”或“detailed”）。将 `<project-root>` 替换为实际的项目根路径。逐字输出代理返回的字符串 — 它包含完全格式化的 ASCII 可视化。 ### 代理提示 > 将 PMM 内存状态渲染为 ASCII 可视化。这是一个只读任务——不要编辑任何文件。您可以运行 git 命令来获取时间戳。 >> **项目根：** `<project-root>` > **深度级别：** `<level>` > > ### 深度级别 > > - `status` — 仅热图 > - `summary` — 热图 + 聚类列表 + 最后 5 个时间线条目 > - `detailed` — 完整 ASCII：图表 + 热图 + 相似度矩阵 + 聚类 > > ### 可视化 1：热图 — 文件活动（所有级别） >> 1. 读取`<project-root>/memory/config.md` 获取活动文件列表 > 2. 对于每个活动文件，运行： `git log -1 --format="%ar|%at" -- memory/<filename>` > 3. 将 unix 时间戳映射到热度级别： > - `████` = 修改 < 5 分钟

更新日志: Source: GitHub https://github.com/NominexHQ/pmm-plugin

目录结构

当前层级: tree/main/local/pmm-dump/

📄 SKILL.md 4.6 KB

SKILL.md

---
name: pmm-dump
description: >
  Dump PMM memory state as ASCII art in the terminal. Three depth levels: status (heatmap
  only), summary (heatmap + clusters + timeline), detailed (full ASCII). Runs as a subagent.
  Use when the user runs /pmm-dump or asks for a text-based memory overview. Trigger on:
  "pmm-dump", "/pmm-dump", "dump memory", "ascii memory", "text memory overview", "show memory
  heatmap", "memory dump", or any request for a text-based ASCII visualization of memory state.
---
# PMM Dump

Render PMM memory state as inline ASCII visualizations. Runs as a subagent to keep the main context clean.

**Depth level:** $ARGUMENTS (default: `status` if empty or not provided)

## Invocation

- `/pmm-dump` or `/pmm-dump status` — heatmap only (status level)
- `/pmm-dump summary` — heatmap + cluster list + last 5 timeline entries
- `/pmm-dump detailed` — full ASCII: graph map + heatmap + similarity matrix + clusters

## Behaviour

Dispatch a `general-purpose` agent using the `Readonly Agent Model` from `memory/config.md` (default: `haiku`) with the prompt below. Replace `<level>` with the depth level (`status`, `summary`, or `detailed`). Replace `<project-root>` with the actual project root path.

Output the agent's returned string verbatim — it contains the fully formatted ASCII visualization.

### Agent Prompt

> Render PMM memory state as ASCII visualizations. This is a READ-ONLY task — do not edit any files. You may run git commands for timestamps.
>
> **Project root:** `<project-root>`
> **Depth level:** `<level>`
>
> ### Depth Levels
>
> - `status` — Heatmap only
> - `summary` — Heatmap + cluster list + last 5 timeline entries
> - `detailed` — Full ASCII: graph map + heatmap + similarity matrix + clusters
>
> ### Visualization 1: Heatmap — File Activity (all levels)
>
> 1. Read `<project-root>/memory/config.md` to get the list of active files
> 2. For each active file, run: `git log -1 --format="%ar|%at" -- memory/<filename>`
> 3. Map the unix timestamp to a heat level:
>    - `████` = modified < 5 minutes ago
>    - `███░` = modified < 30 minutes ago
>    - `██░░` = modified < 2 hours ago
>    - `█░░░` = modified < 24 hours ago
>    - `░░░░` = modified > 24 hours ago or never
> 4. Check if file is empty/template-only (only has comments and headers, no real content)
> 5. Sort by recency (most recent first)
> 6. Render as aligned table with legend
>
> ### Token Burn Estimate (all levels)
>
> After the heatmap, show a token burn estimate:
>
> 1. Count total characters across all `.md` files in `memory/`: `cat memory/*.md | wc -c`
> 2. Estimate read tokens: `total_chars / 4`
> 3. Get last memory diff stats: `git diff HEAD~1 --stat -- memory/ | tail -1` (extract insertions + deletions)
> 4. Estimate write tokens: `(insertions + deletions) * 20 / 4`
> 5. Render:
> ```
> Token Burn (per save)
>   Read:  ~12,400 tok    Write: ~850 tok    Total: ~13,250 tok
> ```
>
> Use compact single-line format for dump. Round to nearest 50. Comma-separate thousands.
>
> ### Visualization 2: Clusters + Timeline (summary and detailed only)
>
> If level is `summary`:
> 1. Read `<project-root>/memory/vectors.md`
> 2. Parse cluster lines: `Cluster: name → [members] | theme: ...`
> 3. Render cluster names with member count
> 4. Read `<project-root>/memory/timeline.md`
> 5. Show last 5 timeline entries
>
> ### Visualization 3: Graph Map (detailed only)
>
> If level is `detailed`:
> 1. Read `<project-root>/memory/graph.md`
> 2. Parse all `[[A]] → relationship → [[B]]` lines
> 3. Group edges by `##` section headers
> 4. Render each group as an ASCII subgraph:
>    - Nodes boxed: `┌──────┐ │ name │ └──────┘`
>    - Edges: `───relationship──▶`
>    - Group under section name
>    - If >15 edges, show top-level structure only (collapse leaf nodes)
>    - If empty/template: "No relationships recorded yet."
>
> ### Visualization 4: Similarity Matrix (detailed only)
>
> If level is `detailed`:
> 1. Read `<project-root>/memory/vectors.md`
> 2. Parse similarity lines: `[[A]] ↔ [[B]] | score: X.XX | basis: ...`
> 3. Parse cluster lines: `Cluster: name → [members] | theme: ...`
> 4. Build sparse similarity matrix
> 5. Render as aligned ASCII table (abbreviate names to 6 chars, `·` for no data)
> 6. Render clusters as tree with `┌├└` chars
>
> ### Output Format
>
> Return the fully formatted ASCII output. Keep it compact. No preamble — start directly with the visualization.

## Notes

- The agent does all file reads and git commands — main context stays clean
- Output the agent's return value verbatim (it's the formatted ASCII visualization)

登录后下载/点赞/收藏 ❤ 6 | ★ 0

pmm-dump

目录结构

SKILL.md

举报内容

提示