pmm-query

Query structured PMM memory files with filters and deep traversal.

15|8|Updated Mar 21, 2026
One-click install
npx skills add https://github.com/NominexHQ/pmm-plugin --skill pmm-query
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: pmm-query
Source: https://github.com/NominexHQ/pmm-plugin/tree/main/local/pmm-query
Command: npx skills add https://github.com/NominexHQ/pmm-plugin --skill pmm-query

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Prevents repetitive questioning and context rot by surfacing relevant decisions, lessons, timelines, and preferences from structured PMM memory files before asking any agent to reason or act.

Core Features & Use Cases

  • Context-aware routing: Uses the built-in table to map user intent to the appropriate memory files with attribution, date, and file-scope filters so queries target decisions, preferences, tone, progress, graphs, timelines, rules, assets, and background knowledge.
  • Deep traversal and fallbacks: Supports the deep flag to expand results through vectors, graph edges, and taxonomy siblings, and escalates to full-file reads or git-history retrieval when in-window memory yields no hits.
  • Flexible output modes: Produces concise, cited prose answers or detailed dumps, respects session start modes, and enriches responses with related graph or asset context only when it adds value.

Quick Start

Ask pmm-query for a summary of the latest decision about pricing rollouts.

Frequently Asked Questions about pmm-query

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I recall past project decisions from structured memory files without losing context?

Structured memory recall from PMM files surfaces relevant decisions, timelines, and preferences using context-aware routing. It maps user intent to appropriate memory files with attribution and date filters, preventing context rot before agent reasoning.

What's the best way to search across git history and vector embeddings for project background knowledge?

Using the deep flag expands memory queries through vector search, graph traversal, and taxonomy siblings. When in-window memory yields no hits, it escalates to full-file reads or git-history retrieval to find background knowledge.

How does memory query routing work when filtering decisions by attribution, dates, and file scope?

Memory query routing uses a built-in table to map user intent to appropriate memory files. It applies attribution, date, and file-scope filters so queries specifically target decisions, preferences, tone, progress, graphs, timelines, rules, and assets.

Can I get a detailed dump of all stored preferences and timelines instead of a summarized answer?

Yes, flexible output modes produce either concise cited prose answers or detailed dumps. The dump mode extracts all matching preferences, timelines, and related graph or asset context directly from structured PMM memory files.

Why does my memory query return incomplete context when searching across multiple files?

Incomplete context recall occurs when in-window memory yields no hits. Applying the deep flag expands results through vectors, graph edges, and taxonomy siblings, and escalates to full-file reads or git-history fallback when necessary.

Do I need pre-existing structured PMM files to use context-aware memory recall?

Yes, context-aware memory recall operates on pre-existing structured PMM memory files. It parses query modifiers to route intent across decisions, preferences, timelines, graphs, and assets stored within your Claude-powered files.