Agent Skills are reusable playbooks you load into an AI agent so it can run a specific post-sales workflow with the same structure every time, and Quivly publishes twenty-one of them for free in the Quivly Skills repo on GitHub for customer success, customer engineering, and other post-sales teams to use in Claude, Cursor, ChatGPT, Quivly, or any Agent Skills–compatible client.
What are Agent Skills?
An Agent Skill is a small folder with a SKILL.md file that tells an agent when to use the skill, which steps to follow, what the output should look like, and which related skills to pull in next, so the agent treats the workflow as a known procedure rather than a one-off chat reply.
Skills follow the open Agent Skills specification, which means the same SKILL.md can travel across tools that support the format, and you can version and share them the way you would share any other playbook in a repo.
What do they do?
The twenty-one skills in customer-engineering/ cover the motions CS and CE teams run every day, week, quarter, and across an account’s lifecycle. Each link below opens that skill’s SKILL.md.
Daily
- Meeting prep — builds a pre-call brief from recent conversations, open tickets, usage changes, and talking points
- Account 360 — assembles a full account snapshot across profile, health, revenue, stakeholders, and risks
- Post-call followup — turns the latest call into owned action items, commitments, and a draft recap
- Ticket pulse — reads support health across aging tickets, recurring themes, and escalation candidates
- Inbox triage — ranks what needs attention today across a book of business
Weekly
- Portfolio digest — reviews health movers, usage droppers, and renewals inside ninety days
- Usage drop investigation — diagnoses why product usage fell before you reach out
- Health drop diagnosis — breaks a health score drop into the components and events behind it
- At-risk report — rolls up at-risk accounts with ARR exposure, risk drivers, and named plays
- Renewal forecast — forecasts next quarter’s renewals with per-account confidence and reasoning
Lifecycle
- Sales handoff — prepares the deal-close handoff with context, contacts, AE commitments, and a first-touch agenda
- Kickoff prep — designs the onboarding kickoff with agenda, stakeholders, milestones, and success criteria
- Onboarding stall — finds where time-to-value is stuck for a new account
- Renewal risk — runs a sixty-to-ninety-day renewal read across value, usage, relationship, and commercial posture
- Churn save plan — builds a root-caused save plan for a red account, with stop conditions
- Champion change — maps exposure and a successor plan when a champion or sponsor leaves
- Escalation brief — drafts an executive one-pager with timeline, stakes, temperature, and talking points
- Customer health review — runs a structured health review with scoring dimensions and an action plan
Quarterly
- QBR prep — prepares the QBR or EBR with a value story, evidence, gaps, forward plan, and renewal positioning
- Expansion scan — surfaces upsell and cross-sell signals from usage pressure, whitespace, and buying language in calls
- Advocacy finder — matches reference and case-study candidates to the advocacy ask they are likely to accept
Skills also point to related skills in their footers, so a health drop can move into diagnosis and then into a churn save plan as one continuous workflow.
Who is it for?
These skills are for CSMs and customer engineers who run meetings, renewals, and saves across a book of accounts, for CS and CE leaders who want the team’s AI workflows to follow the same playbooks, for RevOps and AI platform owners who are evaluating Agent Skills for post-sales, and for builders who want to add new skills through the repo template.
How to use it
You can load the full set with the prompt in the next section, install them in Claude Code with /plugin marketplace add quivly/skills and then pick the skill sets you need, or clone individual skill folders from github.com/quivly/skills into your agent’s skills directory using the client paths documented at agentskills.io.
If you want the skills to pull live customer context while they run, connect the Quivly MCP connector so tool calls resolve against your account data; if you run the skills without connected tools, the agent asks you for the inputs it needs as it goes.
Prompt to use
Copy this into Claude, Cursor, ChatGPT, or any AI chat interface, then let the agent confirm what it loaded before you ask for a specific workflow:
Go to https://github.com/quivly/skills, read the README and the skills under customer-engineering/, then load all of them as Agent Skills I can use. For each skill, follow its SKILL.md when I ask for that workflow. Start by confirming which skills you loaded and briefly what each one does.Once the agent lists the skills, you can ask for something concrete, such as meeting prep for a named account ahead of a date, or a churn save plan for an account that just went red.
Get the repo
The collection lives at github.com/quivly/skills under the MIT License, and you can star, fork, or open a PR when you have a skill the catalog is missing. If you want help wiring the first skill into your stack, grab time with Chandrika.
