agent-memory-systems

Memory is the cornerstone of intelligent agents. Without it, every interaction starts from zero. This skill covers the architecture of agent memory: short-term (context window), long-term (vector s...

Install with Tessl CLI

npx tessl i github:sickn33/antigravity-awesome-skills --skill agent-memory-systems

What are skills?

1.10x

Quality

11%

Does it follow best practices?

Impact

76%

1.10x

Average score across 3 eval scenarios

Optimize this skill with Tessl

npx tessl skill review --optimize ./skills/agent-memory-systems/SKILL.md

SKILL.md

Review

Evals

Evaluation results

56%

Helpdesk Knowledge Retrieval Pipeline

Chunking and retrieval design

Criteria

Without context

With context

Contextual chunk enrichment

Chunk size variation

20%

Metadata pre-filter

100%

Temporal scoring

100%

Embedding model tracking

13%

Retrieval-quality framing

60%

80%

Separate data categories

50%

40%

Ranked result limiting

100%

Without context: $0.5108 · 2m 25s · 18 turns · 19 in / 8,312 out tokens

With context: $0.6035 · 2m 52s · 19 turns · 20 in / 10,239 out tokens

74%

12%

Personal Assistant Memory Module

Memory storage and conflict handling

Criteria

Without context

With context

Selective storage

66%

Conflict detection

90%

100%

Conflict resolution strategy

93%

100%

Memory decay or pruning

93%

100%

Multiple memory categories

13%

No single flat store

38%

Design rationale documented

80%

90%

Without context: $0.3128 · 1m 30s · 12 turns · 12 in / 5,157 out tokens

With context: $0.5379 · 2m 27s · 24 turns · 323 in / 7,986 out tokens

100%

Coding Assistant Context Assembly Module

Token budget and memory type selection

Criteria

Without context

With context

Explicit token budgets

100%

Multiple distinct memory types

100%

Short-term vs long-term separation

100%

Retrieval quality over quantity

50%

100%

Token counting present

100%

Budget overflow handling

100%

Architecture document

100%

Without context: $0.3186 · 1m 31s · 12 turns · 12 in / 5,262 out tokens

With context: $0.5181 · 2m 40s · 18 turns · 269 in / 8,995 out tokens

Evaluated: 8 days ago
Agent: Claude Code
Model: Claude Sonnet 4.6

Table of Contents

Helpdesk Knowledge Retrieval Pipeline Personal Assistant Memory Module Coding Assistant Context Assembly Module

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.