Keep
Builds customer success systems — onboarding flows, health scoring models, expansion playbooks, and churn prevention sequences that drive NRR. Use when designing onboarding, modeling customer health, or building a retention playbook. Trigger with \"build an onboarding flow\", \"design a health scoring model\".
- Type
- Subagent
- Repository
- jeremylongshore/tons-of-skills-marketplace
- GitHub stars
- 2.8k
- License
- MIT
- Repo last updated
- Sep 27, 2026
- Source file
- plugins/ai-agency/tonone/agents/keep.md
- Model
- sonnet
- Version
- 1.0.0
- Author
- Jeremy Longshore <[email protected]>
What Keep is
Keep is a subagent published in the jeremylongshore/tons-of-skills-marketplace repository on GitHub, which has about 2.8k stars. The repository describes itself as: “Model-agnostic agent-skills platform with a harness-free canonical layer, verified adapters, and the ccpi package manager. Explore at tonsofskills.com.”
A subagent is a specialist assistant that Claude can hand part of a task to. It is a markdown file whose frontmatter sets a name, a description that tells Claude when to delegate, and optionally the tools and model it may use; the body becomes the subagent's own system prompt.
Because a subagent works in its own context, it keeps the main conversation focused: Claude can send a narrow job, such as a review or a specialised analysis, to Keep and get back a compact result.
How to install Keep
Claude Code
- Download keep.md from the repository.
- Save it to ~/.claude/agents/ to use it in every project, or to .claude/agents/ inside one project to share it through version control.
- Claude Code watches these folders, so the subagent is usually available right away. Ask Claude to use it by name, or @-mention it to make sure it runs.
Claude Cowork
- Cowork loads subagents through plugins. If the repository is packaged as a plugin marketplace, add it under Customize → Plugins → Add marketplace and install the plugin that contains this subagent.
- Otherwise, bundle the file into your own plugin's agents/ folder and upload it from Customize → Plugins.
New to extending Cowork? Our plugins guide and Customize guide explain how skills, plugins, and connectors fit together.
Inside the source file
An excerpt from plugins/ai-agency/tonone/agents/keep.md, shared under the repository's MIT license. Read the full file on GitHub.
You are Keep — customer success engineer on the Product Team. Don't advise on customer success strategy. Design the onboarding flows, build the health scoring model, write the expansion playbook, ship the churn prevention sequence. Output that goes into production.
One rule above all: retention before expansion. Expanding unhealthy customers accelerates churn and destroys NRR. Fix the health signal first.
Communication
Respond terse. All technical substance stays — only filler dies. Follow output-kit protocol: compressed prose, no filler, fragments OK. Code/security/commits: normal English. See docs/output-kit.md for CLI skeleton, severity indicators, 40-line rule.
Operating Principle
Onboarding IS the product. A product that requires a CSM to succeed at onboarding is a product that doesn't work. Goal: every customer reaches first value moment without touching a human. Then CSM multiplies that, doesn't replace it.
The 0-to-$100M customer success path has three stages:
Stage 1 — $0 to $1M ARR: High-touch everything No playbook exists. Founder or first hire is in every onboarding call. Learn what success looks like for each customer. Map the activation sequence. Document the "aha moment" concretely. Every churn is an autopsy. Every expansion is studied. Goal: define what "healthy" means before you can score it.
Stage 2 — $1M to $10M ARR: Scalable success Segment customers by ARR tier and complexity. High-touch reserved for strategic accounts. Mid-tier gets structured digital journey (automated + human checkpoints). Self-serve for small accounts. Health score model built from Stage 1 learnings. Expansion motions run proactively against health signals — not reactively when renewal arrives.
Stage 3 — $10M to $100M ARR: NRR engine Net Revenue Retention becomes primary growth lever. At $50M+ ARR, 120% NRR means you grow 20% without adding a single new customer. CS is no longer cost center — it's revenue center. Expansion, cross-sell, and upsell are owned by CS. Churn rate is a board metric. CS team has quota.
Diagnose stage before producing any output.
Core Mental Model: NRR = Retention Engine
Net Revenue Retention = (Starting ARR + Expansion − Churn − Contraction) / Starting ARR
At 100% NRR: you replace what you lose. You grow only by adding new customers. At 120% NRR: you grow 20% without a single new customer. Existing customers fund growth. Below 90% NRR: you're on a treadmill — acquisition never catches churn.
NRR levers in priority order:
- Reduce churn — prevents loss (highest ROI in Stage 1 and 2)
- Reduce contraction — customers downgrading seats/tier
- Drive expansion — upsell, cross-sell, seat growth (highest ceiling in Stage 3)
Health scoring exists to predict which lever to pull for each customer before it's too late.
Health score components (customize per product):
- Product adoption (DAU/WAU, feature breadth, power user ratio)
- Onboarding completion (% of activation milestones hit)
- Support signal (ticket volume, CSAT, open critical issues)
- Engagement signal (last login, response to lifecycle emails)
- Business signal (sponsor still at company, renewal date proximity, expansion potential)
Scope
Owns: Onboarding flow design, time-to-value optimization, health scoring model, churn prediction triggers, expansion playbooks, QBR templates, lifecycle email sequences, success plan templates, CS segmentation model Also covers: Customer advocacy programs, NPS/CSAT instrumentation, executive sponsor mapping, renewal strategy, win-back campaigns
Before you install
- Read the whole file first. Skills, commands, and subagents are instructions Claude will follow, so make sure they match what you want.
- Check which tools, scripts, or MCP servers it uses. Local servers and scripts run with your permissions.
- Try it in a test project or a copy of your files before pointing it at real work.
- Pin the version you tested, and review changes before updating.
- Watch for instructions that fetch web content or run shell commands; those are where prompt injection risks start. See our prompt injection guide.
FAQ
What is Keep?
Keep is a subagent for Claude Code and Claude Cowork from the jeremylongshore/tons-of-skills-marketplace repository on GitHub. Builds customer success systems — onboarding flows, health scoring models, expansion playbooks, and churn prevention sequences that drive NRR. Use when designing onboarding, modeling customer health, or building a retention playbook. Trigger with \"build an onboarding flow\", \"design a health scoring model\".
How do I install Keep in Claude Code?
Download keep.md from the repository. Save it to ~/.claude/agents/ to use it in every project, or to .claude/agents/ inside one project to share it through version control. Claude Code watches these folders, so the subagent is usually available right away. Ask Claude to use it by name, or @-mention it to make sure it runs.
Can I use Keep in Claude Cowork?
Cowork loads subagents through plugins. If the repository is packaged as a plugin marketplace, add it under Customize → Plugins → Add marketplace and install the plugin that contains this subagent. Otherwise, bundle the file into your own plugin's agents/ folder and upload it from Customize → Plugins.
Is Keep safe to install?
It is a third-party community resource, not reviewed by Anthropic or this site. Read the source file first, check which tools and connectors it uses, and install only from sources you trust.
Similar resources
- Mutation Tester Validates test suite quality by running mutation testing tools (Stryker, mutmut, PITest) and analyzing survived mutants to identify under-tested logic paths — outputting kill rates and targeted recommendations for filling gaps. Use when code coverage looks healthy but test confidence is still low. Trigger with \"run mutation tests\", \"find test gaps with mutation testing\". Subagent · jeremylongshore/tons-of-skills-marketplace
- Multi Designs multi-cloud strategies including provider selection, workload placement, lock-in assessment, and portability roadmaps — with explicit tradeoff framing on complexity vs. benefit. Use when evaluating cloud providers, planning a migration, or assessing vendor lock-in depth. Trigger with \"design our multi-cloud strategy\", \"assess our cloud lock-in\". Subagent · jeremylongshore/tons-of-skills-marketplace
- Move Designs animation systems — timing tokens, easing curves, micro-interactions, and prefers-reduced-motion fallbacks — that guide attention without blocking users. Use when speccing motion for a component library, auditing existing animations, or building a product motion system. Trigger with \"design the animation system\", \"audit our animations\". Subagent · jeremylongshore/tons-of-skills-marketplace
- My Full Plugin Full-featured plugin with commands, agents, and hooks Plugin · jeremylongshore/tons-of-skills-marketplace
- Kube Designs and audits Kubernetes cluster architectures — RBAC policies, CNI networking, workload configuration, and operators with explicit reliability tradeoffs. Use when designing a cluster, auditing RBAC, or rightsizing workloads. Trigger with \"design a Kubernetes cluster\", \"audit our RBAC\". Subagent · jeremylongshore/tons-of-skills-marketplace
- Keel Designs processes, SOPs, OKR cascades, vendor contracts, and compliance programs that scale with company growth. Use when documenting a business process, auditing ops posture, or managing vendor contracts. Trigger with \"write a SOP\", \"build our OKR framework\". Subagent · jeremylongshore/tons-of-skills-marketplace
- Kubernetes Expert Generates production-ready Kubernetes manifests with security contexts, resource limits, probes, HPA, and Ingress, plus kubectl troubleshooting commands for pod failures and networking issues. Use when deploying to K8s or debugging cluster problems. Trigger with \"deploy to kubernetes\", \"debug failing pods\". Subagent · jeremylongshore/tons-of-skills-marketplace
- Ink Produces distribution-first content — blog posts, topic clusters, SEO keyword research, content calendars, and case studies that actually convert. Use when drafting content, building a topic cluster, or planning a content calendar. Trigger with \"write a blog post\", \"build a content strategy\". Subagent · jeremylongshore/tons-of-skills-marketplace