skills.rest
jayho-k
jayho-kCOMMUNITY
Explore all skills by this author →
GitHub

context-loader

Searches and injects relevant memories into context before starting a task.

1|Updated Jan 13, 2022
Search Tags:#context#search#agent#retrieval#memory#session#mem0
One-click install
npx skills add https://github.com/jayho-k/TIL --skill context-loader-jayho-k
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: context-loader
Source: https://github.com/jayho-k/TIL/tree/main/AI/mem0/code/mem0_code_analize/mem0/integrations/pi-agent-plugin/skills/context-loader
Command: npx skills add https://github.com/jayho-k/TIL --skill context-loader-jayho-k

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? When starting a new task or switching topics, an AI agent lacks awareness of past decisions, preferences, and lessons stored in long-term memory, forcing users to repeat context manually. ## Core Features & Use Cases - Parallel Memory Search: Runs 2-4 parallel queries against the mem0 memory store using different angles such as topic, people, and project references. - Deduplicated Context Block: Merges results by memory ID and outputs a compact block of up to 10 memories tagged by type (decisions, preferences, lessons). - Silent-on-Empty Behavior: Produces no output when no relevant memories exist, avoiding noise in the conversation. - Use Case: When a user says "what do we know about project X", the skill searches mem0 for related decisions and preferences and injects a summarized context block before work begins. ## Quick Start Ask the agent to load context for a topic, for example: "Load context for the payments migration project before we start planning."

Frequently Asked Questions about context-loader

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

FAQPage Schema
How do I load past memories into an AI agent's context?▼

Run parallel search queries against the mem0 memory store using topic, people, and project angles, then deduplicate results by memory ID. The skill outputs a compact block of up to 10 relevant memories tagged as decisions, preferences, or lessons.

How to search mem0 memories before starting a task?▼

Use the mem0_memory tool with action set to search and issue 2-4 queries from different angles such as subject name, mentioned people, and project references. Merge the results and keep only the most relevant entries.

Can context-loader modify or delete stored memories?▼

No, the skill is strictly read-only. It only performs search operations and never modifies or deletes memories in the mem0 store.

What happens when no relevant memories are found?▼

The skill outputs nothing when search returns zero results. It stays silent on empty context rather than announcing that no memories exist.

How many memories can be injected into context at once?▼

A maximum of 10 memories are returned, keeping only the most relevant results. Memories already visible in the current session context are skipped to avoid duplication.

Related Skills

Explore More Software Engineering Skills→
adobeadobe
1.5k

session-retrospective

Capture session corrections and lessons into .ai/memory/lessons.md.

Official
Advanced
hiddink-aihiddink-ai
1

memory-save

Save session context with tasks, decisions, and code changes to claude-mem.

Official
Intermediate
miriamdongmiriamdong
2

consolidate-memory

Merge duplicate memory entries and prune outdated facts from MEMORY.md.

Community
Intermediate
ruvnetruvnet
70.1k

embeddings

Generates vector embeddings with HNSW indexing for semantic search and similarity queries.

Community
Intermediate