Best Agent Setup and Memory Plugins, Skills, and MCP Servers in 2026

AllSetupSkillsBehaviorMemory

#	Name	Best for	Est. installs About install estimates
1	Find Skills	Discovering skills you can install	1.4M
2	Grill Me	Grilling a plan until decisions are explicit	670K
3	Caveman	Terse replies, fewer output tokens	410K
4	Skill Creator	Building and measuring your own skills	390K
5	CLAUDE.md Management	CLAUDE.md files that stay current	290K
6	Writing Great Skills	A method for authoring skills	240K
7	Claude Code Setup	A shortlist of automations for your repo	230K
8	claude-mem	Session memory that persists locally	220K
9	Graphiti	Temporal knowledge-graph memory	100K
10	Mem0	Portable long-term agent memory	100K
11	Remember	Local session-to-session handoffs	70K
12	Learning Output Style	Learning while Claude codes	60K
13	Interview Me	Interviewing you before it starts	45K
14	Ask Questions If Underspecified	A few must-ask questions, then go	25K

Show moreShow less

---

​

Find Skills

by Vercel Labs

Discovering skills you can install

1.4Mestimated installs

What it isFind Skills teaches an agent to search the open skill ecosystem and pull in relevant packages on demand. It rides on the shared Skills CLI and skills.sh catalog rather than being the catalog itself, acting as the discovery layer.

When to useReach for it when a user needs a capability that likely already exists as a skill. It ranks candidates by popularity and publisher signal, but discovery is not a safety check. Every package it surfaces carries its own code and permissions, so review before installing.

InstallSkill

​

Grill Me

Grilling a plan until decisions are explicit

670Kestimated installs

What it isGrill Me makes an agent investigate the discoverable facts itself, then ask you one decision question at a time until a plan or idea is genuinely understood. The visible command is a thin wrapper around the collection’s reusable Grilling skill.

When to useReach for it on non-code plans, product decisions, or genuinely ambiguous work where a wrong assumption is expensive, and it keeps you in control of the real decisions. Skip it for clear, reversible tasks; for codebase design the publisher points to Grill with Docs, which also preserves domain language.

InstallPluginSkill

​

Caveman

Terse replies, fewer output tokens

410Kestimated installs

What it isCaveman forces a coding agent into deliberately terse replies, stripping the conversational padding while leaving code, commands, and error output intact. The logic lives entirely in the extension, with no external service behind it.

When to useIt earns its place when an agent’s explanations are consistently longer than you need, cutting reading time and trimming output tokens. The trade-off is blunt: it is a poor fit for teaching, stakeholder-facing writing, or any work where visible reasoning is the point.

InstallPluginSkill

​

Skill Creator

by Anthropic

Building and measuring your own skills

390Kestimated installs

What it isSkill Creator is a toolkit for building and refining Claude skills across four modes: Create, Eval, Improve, and Benchmark. Instead of writing a skill and hoping it works, you get structured stages for drafting, testing, and measuring it.

When to useThe payoff is measurement: graders, blind comparisons, validation, and benchmark reports let you see whether a change actually improved a skill rather than guessing. Worth knowing that repeated evaluation runs consume real model usage and execute local scripts, so heavy iteration has a cost.

InstallPluginSkill

​

CLAUDE.md Management

by Anthropic

CLAUDE.md files that stay current

290Kestimated installs

What it isCLAUDE.md Management is a project-memory workflow with two moves. An audit skill checks your CLAUDE.md files against the current codebase, and a session command turns durable learnings from a session into proposed diffs you review before they land.

When to useIt helps when project instructions drift after code changes, or when a session surfaces a durable command or gotcha worth keeping. It aims to keep instructions useful rather than bloated. Treat its scores and diffs as prompts for judgment, and keep secrets, ephemeral state, and verbose summaries out.

What you needNothing beyond Claude - currently a Claude-only plugin.

InstallPlugin

​

Writing Great Skills

by Matt Pocock

A method for authoring skills

240Kestimated installs

What it isA community meta-skill from Matt Pocock for designing, writing, and improving agent skills - a repeatable method for authoring skill instructions and judging whether one is clear and genuinely useful.

When to useWhen you are creating or revising a skill and want more than a blank file. Despite the name it operates on the skill itself, not prose; it pairs with Skill Creator, which adds the evaluation and benchmarking loop.

InstallSkill

​

Claude Code Setup

by Anthropic

A shortlist of automations for your repo

230Kestimated installs

What it isClaude Code Setup is a read-only recommender that scans your repository and suggests a short list of automations to add: MCP servers, skills, hooks, subagents, and slash commands. It reads the project but changes nothing on its own.

When to useIt helps most when the extension ecosystem feels overwhelming, because it narrows the field to a shortlist tied to your project’s structure, dependencies, and patterns rather than a generic popularity list. Since it only suggests, verify the source, maintenance, permissions, and compatibility of anything third-party before you add it.

What you needNothing beyond Claude - currently a Claude-only plugin.

InstallPlugin

​

claude-mem

Session memory that persists locally

220Kestimated installs

What it isclaude-mem is a memory layer that watches an agent session, summarizes what happened through an AI model, and files it in local SQLite and vector indexes. Later sessions can pull back the relevant pieces. The MCP search tools are one part of this larger system.

When to useIt shines on long projects where sessions keep losing decisions, and it retrieves history progressively instead of dumping everything into context. The catch: it records substantial session activity and sends it to your chosen model, so vet retention, privacy exclusions, and cloud sync before pointing it at sensitive repositories.

What you needIt runs a local memory service; summarization calls a model provider.

InstallPluginCLI

​

Graphiti

by Zep

Temporal knowledge-graph memory

100Kestimated installs

What it isGraphiti gives an agent a temporal knowledge graph for long-term memory - episodes, entities, relationships, and facts that change over time - queried with hybrid, time-aware search. The agent route is Zep’s official but explicitly experimental MCP server, running on the broader Graphiti library and a graph database.

When to useWhen flat memory notes are not enough and you want provenance and relationships an agent can reason over, and your team can operate the infrastructure. It is the most capable memory option here, and the most demanding to run and secure.

What you needA local graph database (FalkorDB or Neo4j), a model or embedding API key, and a Docker or Python runtime; use MCP 1.0.2 or later.

InstallMCP

​

Mem0

Portable long-term agent memory

100Kestimated installs

What it isMem0 is a persistent memory layer for agents and AI applications, reachable through a hosted MCP server, plugins with lifecycle hooks, portable skills, SDKs, or a self-hostable open-source library. Agents can explicitly add, search, update, and delete memories.

When to useThe simple path gives an agent direct memory commands; the plugin route goes further, pulling context at task start and capturing learnings before they scroll away. The hosted free tier allows 10,000 additions and 1,000 retrievals monthly. Because automatic capture can store sensitive context, set retention and access rules first.

What you needA Mem0 account and API key for the hosted route; self-hosting is the alternative.

InstallPluginPluginMCP

​

Remember

Local session-to-session handoffs

70Kestimated installs

What it isA community continuity plugin for Claude Code that extracts and compresses each session into persistent daily memory and handoff files, so project context survives across separate sessions without a hosted service.

When to useWhen you want local, file-based continuity rather than sending session activity to a cloud memory provider. It overlaps with claude-mem and Mem0 but keeps everything on disk - review what its hooks persist.

What you needClaude Code, where it runs locally through session hooks; no verified Codex or Cursor route.

InstallPlugin

​

Learning Output Style

by Anthropic

Learning while Claude codes

60Kestimated installs

What it isA first-party Claude behavior plugin that changes how Claude works a coding task: it explains the reasoning behind implementation choices and pauses at decision points for you to write small, meaningful pieces of the code yourself.

When to useWhen understanding and practice matter more than the fastest possible completion. The trade-off is direct: the added explanations and hand-offs cost tokens and time, so it is a poor fit for routine production work or urgent fixes.

What you needNothing beyond Claude - currently a Claude-only plugin.

InstallPlugin

​

Interview Me

by Addy Osmani

Interviewing you before it starts

45Kestimated installs

What it isA community skill that has the agent interview you one focused question at a time about an ambiguous task, continuing until it is roughly 95% confident, then restating the work as a clear specification before starting.

When to useBefore complex or high-stakes work where requirements are incomplete and a wrong assumption is expensive. For simple tasks it is overkill - and note the individual skill is the install to want, not the 24-skill collection it ships in.

InstallSkill

​

Ask Questions If Underspecified

by Trail of Bits

A few must-ask questions, then go

25Kestimated installs

What it isTrail of Bits’ focused behavior plugin: the agent spots material ambiguity, asks a small set of must-have questions - roughly one to five - and pauses until they are answered or you approve reasonable assumptions.

When to useWhen a few unanswered requirements could change the implementation but you do not want a full discovery interview. It is the lighter, more bounded cousin of Interview Me; skip it for routine tasks or when you have already authorized assumptions.

InstallPluginPluginSkill

​

Choose by use case

Deciding what to add, or keeping your setup honest as the code moves?

→ Claude Code Setup - a read-only repo scan that narrows the ecosystem to a shortlist tied to your project (Claude-only)

→ CLAUDE.md Management - audits your instructions against the current code and proposes diffs you review (Claude-only)

Finding or building skills?

→ Find Skills - searches the open skill ecosystem and pulls in packages on demand (review each before installing)

→ Skill Creator - create, evaluate, improve, and benchmark your own skills across four modes

→ Writing Great Skills - a repeatable method for authoring a single skill well

Want the agent to pin down requirements before it starts?

→ Grill Me - for plans and product decisions: it researches what it can, then asks one decision question at a time

→ Interview Me - a thorough one-question-at-a-time interview that restates the work as a spec before starting

→ Ask Questions If Underspecified - a lighter must-ask set, then it proceeds on reasonable assumptions

These overlap - pick one clarification style rather than stacking them.

Changing how the agent explains and hands off work?

→ Caveman - deliberately terse replies; code, commands, and errors stay intact

→ Learning Output Style - explains the reasoning and pauses for you to write key code yourself (Claude-only)

Agent forgetting what past sessions learned?

→ claude-mem - automatic local memory: summarize, store, retrieve progressively

→ Mem0 - explicit memory commands, hosted or self-hosted, portable across apps

→ Remember - local, file-based session handoffs with no hosted service (Claude-only)

→ Graphiti - a temporal knowledge graph for provenance and relationships (heavier setup)

A searchable knowledge store is not agent memory - these preserve working continuity across sessions.

​

Related categories