customer-service
Skill0 starsChanged 2 months agoMITView on GitHub
Manage support chats, emails, or forms for warranty, repair, return, refund, billing, account, or order problems. Use when the assistant should prepare or handle a customer-service interaction. Do not use for ordinary casual conversations.
4 other projects use a version of this.
Adopted by
Most recently taken up by
- pschlieker/getbacklanetheir own version, first seen Nov 8, 2026
- scayver/marketing-skillstheir own version, first seen Oct 13, 2026
- murphye/agent-skills-customer-servicetheir own version, first seen Feb 18, 2026
- vivy-yi/xiaohongshu-skillstheir own version, first seen Jan 22, 2026
Other versions4 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 4
- vivy-yi/xiaohongshu-skills★ 455same name
- scayver/marketing-skills★ 8same name
- murphye/agent-skills-customer-service★ 4same name
- pschlieker/getbacklane★ 1same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository7 other files it gives its agents.
The file
Show its rules and text
We split this skill into the rules it gives an agent
14 rules. 0 are common, 14 are distinctive, and 0 are very specific to this project.
0
Common
Most of the 3 edited versions we read keep these.
14
Distinctive
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 13
Use available context and safe, reversible defaults before asking for input.
- Distinctive: line 28
Keep internal monitoring separate from external messaging.
- Distinctive: line 33
When the agent says "please wait," wait silently for a reasonable human interval, normally at least five minutes for account or order …
- Distinctive: line 35
If a substantive message has been seen but unanswered, wait at least three minutes before one concise follow-up.
- Distinctive: line 38
If the chat displays an imminent inactivity timeout, send one minimal connection message.
- Distinctive: line 45
Before a consequential choice, clarify only terms that could materially change the decision, such as total cost, return obligation and …
- Distinctive: line 48
If the user explicitly says to accept or proceed despite optional unanswered details, stop negotiating and carry out that choice.
- Distinctive: line 54
Do not opt the user into marketing unless they explicitly request it.
- Distinctive: line 58
Stop external messaging once the requested action is confirmed as submitted or processing and the next step, delivery channel, or expected …
- Distinctive: line 60
Do not delay closure solely to obtain an optional case or order reference.
- Distinctive: line 62
Treat the user's direction as the controlling stop rule:
- Distinctive: line 68
If the agent asks whether anything else is needed and the goal is met, close the interaction instead of inventing additional requests.
- Distinctive: line 75
Never present an earlier state as a later one.
- Distinctive: line 77
Keep support messages calm, direct, and normally one compact paragraph.
“Common” is measured: kept by at least 6 in 10 of the 3 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Customer Service
Resolve the user's issue with the fewest necessary external messages. Preserve the user's chosen outcome, authorization boundaries, and preferred pace.
Keep momentum
Use available context and safe, reversible defaults before asking for input. Bundle non-blocking questions and continue useful work. Interrupt only when a missing choice would change the result, privacy, cost, authority, safety, or something hard to undo. Never repeat an answered question.
Define success before messaging
- Identify the requested outcome and the minimum evidence needed to establish it.
- Separate essential decision blockers from useful but optional details.
- Bundle the relevant facts, evidence, and essential questions into one concise message instead of sending them piecemeal.
- Share only information and files authorized for that support destination.
Use human-paced chat behavior
- Keep internal monitoring separate from external messaging. Refreshing or reading the chat does not justify sending a message.
- Send one message at a time. Do not repeat a point while the prior message is unread, while the agent is typing, or after the agent says they are processing the request.
- When the agent says "please wait," wait silently for a reasonable human interval, normally at least five minutes for account or order processing.
- If a substantive message has been seen but unanswered, wait at least three minutes before one concise follow-up. Send no second follow-up on the same point unless the user explicitly asks to escalate.
- If the chat displays an imminent inactivity timeout, send one minimal connection message. Do not use repeated "still connected" nudges otherwise.
- Report to the user only material state changes or a requested status update; do not narrate every poll or minute of waiting.
Limit questions and negotiation
- Before a consequential choice, clarify only terms that could materially change the decision, such as total cost, return obligation and deadline, replacement condition, recurring terms, or a new legal commitment.
- If the user explicitly says to accept or proceed despite optional unanswered details, stop negotiating and carry out that choice. Pause only for a newly disclosed charge, binding term, or materially different offer.
- State corrections to receipts, purchase condition, or account records once clearly and attach the supporting evidence. Do not demand repeated acknowledgements unless the correction is essential and remains disputed.
- Do not opt the user into marketing unless they explicitly request it.
Apply explicit stopping conditions
- Stop external messaging once the requested action is confirmed as submitted or processing and the next step, delivery channel, or expected timeline is known.
- Do not delay closure solely to obtain an optional case or order reference. Record one if the agent provides it.
- Treat the user's direction as the controlling stop rule:
- "Wait" means send nothing and monitor silently only if still useful.
- "Just accept" means accept the identified offer, obtain only the essential confirmation, and stop.
- "End," "stop," or "close the conversation" means send one brief sign-off, close the chat surface, and perform no further follow-up.
- If the agent asks whether anything else is needed and the goal is met, close the interaction instead of inventing additional requests.
Communicate status precisely
Distinguish states such as evidence attached, request accepted, order processing, label emailed, item received, replacement shipped, and replacement delivered. Never present an earlier state as a later one.
Keep support messages calm, direct, and normally one compact paragraph. Keep the user handoff equally concise: current state, their next action, any deadline, and the reference number if already available.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

