[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7EC0ACE0YEV3M2RKP1D7HZJ":3,"skill-same:01M7EC0ACE0YEV3M2RKP1D7HZJ":76},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":40,"commitSha":41,"skillDirectory":42,"forksCount":43,"repositoryHtmlUrl":44,"repositoryHomepageUrl":6,"ownerHtmlUrl":45,"ownerBlogUrl":6,"ownerDescription":46,"updatedAt":47,"install":48,"resources":61},"01M7EC0ACE0YEV3M2RKP1D7HZJ","customer-support",null,"Use when a live support ticket needs triage (intent, P1–P4 priority, SLA clock), an on-voice reply, not a raw template, de-escalation of an angry or churn-risk customer, or escalation with a handoff packet. NOT the help-center article it links to (that is `technical-writing`), a win-back program (`retention`), or the auto-answer bot (`chatbot`).","lza6\u002FArmor-piercing-system-prompts","lza6","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F128880206?v=4","01M7EB3K7WPX5T28EJ2NVHE6EG","Chinese-tingfeng","01M7EB3K97E1KQ16QWKJMFNGX8",11,"MIT","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F03-逆向工程\u002Fcustomer-support\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: customer-support\ndescription: \"Use when a live support ticket needs triage (intent, P1–P4 priority, SLA clock), an on-voice reply, not a raw template, de-escalation of an angry or churn-risk customer, or escalation with a handoff packet. NOT the help-center article it links to (that is `technical-writing`), a win-back program (`retention`), or the auto-answer bot (`chatbot`).\"\ntags: [customer-support, helpdesk, ticket-triage, sla, escalation, macros, de-escalation, csat]\nrecommends: [technical-writing, client-onboarding, retention, chatbot, rag, brand-voice]\norigin: risco\n---\n\n# Customer Support — Run the Desk on a Live Ticket\n\nYou own the **support desk as an operation**: take an incoming ticket, assign priority and intent, start the SLA clock, draft the reply (macro or bespoke), enforce tone, and decide escalation with a clean handoff. The output is a triaged, answered ticket — human judgment, not a code artifact.\n\n## The triage-to-reply loop (the spine)\n\nRun every ticket through these six steps in order. Skipping triage to \"just reply fast\" is how you send the wrong tone to a P1.\n\n```text\n1 Classify intent   →  what does the customer actually want? (bug, how-to, refund, outage, rage)\n2 Set priority      →  P1–P4 (impact × urgency × account tier)\n3 Start SLA clock   →  pick the FRT target NOW, before drafting\n4 Draft reply       →  macro skeleton or bespoke; one personalized line minimum\n5 Tone \u002F voice check →  on-brand, de-escalated, no banned phrases\n6 Escalate or send  →  trigger met? hand off with the packet. else send + log.\n```\n\n## Priority, SLA, and the escalation call\n\nThe branch is at step 6. This table sets priority, both SLA clocks, and whether the ticket leaves your hands:\n\n| Signal in the ticket | Priority | First response (FRT) | Resolution target | Escalate? |\n|---|---|---|---|---|\n| Production down \u002F data loss \u002F payments broken \u002F many users blocked | P1 critical | 15–30 min | ASAP, status updates every 30–60 min | Yes → tier-2\u002Feng + the packet |\n| Enterprise\u002FVIP account, or \"cancel \u002F refund \u002F lawyer \u002F unacceptable\" | P1–P2 | 30 min – 1 h | Same business day | Yes → owner + tier-2 |\n| One user blocked, no workaround | P2 high | 1–2 h | Same business day | Only if unresolved past target |\n| Question with a workaround, billing query | P3 medium | 4–8 h | 1–2 business days | No, unless policy claim is unverifiable |\n| FAQ, cosmetic, feature request | P4 low | ~1 business day | Best effort | No |\n\n**Why pick the FRT target before drafting:** the SLA decides your tone and length. A P1 gets a two-line acknowledgment in 20 minutes, not a polished essay in two hours.\n\nUse **3–4 priority tiers, never a flat SLA.** Too many tiers confuse the desk; too few don't differentiate the outage from the typo (Emailmeter \u002F Hiver, 2026-06-02).\n\n**Acknowledge fast even when you cannot resolve fast.** First-response and resolution are *separate, linked* SLAs: FRT sets the expectation, resolution is the outcome. 89% of customers feel valued by a fast first response even when the full fix takes longer (getMonetizely \u002F Hiver, 2026-06-02). So: acknowledge inside the FRT window with an honest \"here's what I know and when I'll update you,\" then resolve honestly. Never go silent to wait for a perfect answer.\n\nFull P1–P4 matrix by channel (email\u002Fchat\u002Fphone) lives in [`references\u002Fmacros-and-sla.md`](references\u002Fmacros-and-sla.md).\n\n## Macros done right\n\nA macro with **one genuinely personalized sentence raises CSAT; a verbatim template lowers it.** Macros cut first-response time 30–50%, but sent raw they read as scripted (TextExpander \u002F Gorgias, 2026-06-02). So every macro is a skeleton you finish, not a message you forward.\n\nThe 4-beat skeleton — every reply hits these in order:\n\n```text\n1 Empathy \u002F ack   →  name the problem in their words (1 line)\n2 Answer          →  the fix, the status, or the next concrete step\n3 Next step       →  what happens now + when they'll hear back\n4 Close           →  human sign-off, door left open\n```\n\nRules:\n\n- **Fill dynamic fields, then add one bespoke line.** Placeholders like `{{first_name}}`, `{{order_id}}`, `{{ticket_id}}`, `{{date}}` are table stakes — they are not personalization. The personalized line references *this* ticket's specifics.\n- **Never send a macro with a visible `{{placeholder}}`.** A leaked `{{first_name}}` is worse than no name.\n- **One reply, one ask.** If you need three things from the customer, you'll get one. Batch the request or sequence it.\n\nBad → Good (info-request macro):\n\n```text\nBad (verbatim template, robotic):\n  Dear Customer, Thank you for contacting support. We have received your\n  request. In order to assist you, please provide additional information.\n  Your satisfaction is important to us. Regards, The Support Team.\n\nGood (skeleton + one personalized line):\n  Hi Marta — sorry the export keeps failing on you. To dig in I need one\n  thing: the rough time you last hit \"Export CSV\" today (so I can pull the\n  matching server log). I noticed your account is on the new billing plan,\n  so this might be the same migration glitch we patched yesterday — I'll\n  confirm and get back to you within 2 hours. — Eric\n```\n\nThe Good version names the feature, gives a real next step + a time, and the \"new billing plan \u002F migration glitch\" line proves a human read the ticket.\n\nFull skeleton library (ack, info-request, resolved, refund, known-bug, outage, churn-save, escalation-handoff, feature-request decline, gentle-close) is in [`references\u002Fmacros-and-sla.md`](references\u002Fmacros-and-sla.md).\n\n## Tone and de-escalation\n\nApply the brand voice; do not invent it here. The voice guide is authored by [`..\u002Fbrand-voice\u002FSKILL.md`](..\u002Fbrand-voice\u002FSKILL.md) — this skill reads it and writes the reply against it.\n\nFor an angry or churn-risk ticket, de-escalation moves in order:\n\n1. **Acknowledge the feeling before the facts.** \"Three days without a reply is not okay, and I get why you're frustrated\" lands before any fix.\n2. **Take ownership in the first person.** \"I dropped this\" beats \"the team was unable to.\" No passive voice to hide behind.\n3. **One concrete next step with a time.** Vague reassurance reads as a brush-off; \"I'll have an answer by 3pm\" is trust.\n4. **Match their stakes, not their volume.** Stay calm and specific. Do not argue, and never match anger with anger.\n\nBanned phrases — they escalate, not de-escalate:\n\n| Never write | Why it backfires | Write instead |\n|---|---|---|\n| \"Please calm down.\" | Tells them their feeling is the problem. | \"You're right to be frustrated — let's fix it.\" |\n| \"As I said \u002F as previously mentioned\" | Scolds them for not reading. | Just restate it plainly. |\n| \"Per our policy.\" (naked) | Hides behind rules, no human. | Explain the *why*, then the policy. |\n| \"Unfortunately, there's nothing we can do.\" | Dead end, zero path. | Offer the nearest thing you *can* do. |\n| \"This is a known issue.\" (alone) | Sounds like \"we knew and didn't care.\" | Add the status + the ETA or the workaround. |\n\nChecklist before send: feeling acknowledged · first-person ownership · one timed next step · no banned phrase · no `{{placeholder}}` left.\n\nDe-escalating *this* ticket is the job here; running a win-back or renewal **program** is `retention`.\n\n## Escalation\n\nEscalate on **explicit conditions, not vibes.** Standard tiers: T1 frontline\u002Fcommon, T2 specialist\u002Fcomplex, T3 engineering or exec\u002Fcritical (Kapture \u002F Hiver, 2026-06-02). Hand off when any trigger fires:\n\n- **Negative-language signals:** \"unacceptable\", \"cancel\", \"refund\", \"lawyer\", \"chargeback\".\n- **SLA at risk:** the FRT or resolution target is about to blow.\n- **Account tier:** enterprise \u002F VIP.\n- **Confidence below threshold:** if an assisted\u002FAI draft is below ~85% confidence on a policy or price claim, hand to a human — don't guess (swiftflutter \u002F Kapture, 2026-06-02).\n\n**Context must travel with the escalation.** 70% of customers expect the next agent to know their history; making them repeat themselves adds 90–180 seconds per ticket (Forrester via Fini Labs, 2026-06-02). Never escalate a bare ticket link. Attach the handoff packet:\n\n```text\nHANDOFF PACKET\n- Customer + account tier:  Marta R. — Enterprise (VIP)\n- Ticket \u002F priority:        #4821 — P1\n- One-line summary:         CSV export 500s since ~09:00; blocking month-end close.\n- Steps already tried:      Confirmed not browser\u002Fcache; reproduced on staging.\n- Customer sentiment:       Angry, has mentioned \"cancel\" once.\n- Collected variables:      plan=new-billing, last_export=09:14, region=eu-west\n- What I need from you:      Confirm if the billing-migration patch covers this.\n```\n\nThe full escalation matrix (trigger → tier → owner) is in [`references\u002Fmacros-and-sla.md`](references\u002Fmacros-and-sla.md).\n\n## FAQ and KB answering\n\n**Quote the KB or escalate — never invent policy, price, or an ETA.** Hallucinated support answers produced 150+ documented legal cases by mid-2025 (cmswire \u002F swiftflutter, 2026-06-02). The guardrail:\n\n- If the answer is in the KB, **quote it and link the article.** Don't paraphrase a refund window or a price from memory.\n- If it is not in the KB, say \"let me confirm\" and escalate — do not fill the gap with a plausible guess.\n- Deflection is **bounded by KB quality.** Triage-only without a strong KB deflects ~20–30%; well-organized KBs push assisted\u002Fauto resolution to 50–80% (Zendesk \u002F Intercom \u002F Fini Labs, 2026-06-02). A bad answer fast is worse than a correct answer slightly slower.\n\nThis skill *consumes* the KB; it never writes it. If the article the customer needs does not exist, flag the gap — authoring it is a `technical-writing` job, and building the autonomous auto-answer bot on top of the KB is `chatbot` + `rag`.\n\n## Metrics — what to watch and when\n\nNo single metric is sufficient; track the trio plus resolution rate (Armatis \u002F Giva, 2026-06-02):\n\n| Metric | What it measures | When to optimize for it |\n|---|---|---|\n| **CSAT** | Transactional satisfaction; `positive ÷ total × 100` | After any single ticket; the default health signal |\n| **CES** | Customer effort, 1–7 scale | When tickets resolve but customers still churn — friction is the problem |\n| **NPS** | Relationship; promoters − detractors | Quarterly \u002F relationship level, not per-ticket |\n| **FCR \u002F one-touch** | Resolved in one interaction | When repeat contacts are climbing — optimize the first reply |\n\nRule of thumb: optimize **FRT** for reassurance (the customer feeling handled), **FCR** for effort (fewer round-trips). They pull in different directions — don't chase both blindly.\n\n## Anti-patterns\n\n| Anti-pattern | Why it fails | Do instead |\n|---|---|---|\n| Send the macro verbatim | Reads as scripted, drops CSAT | Skeleton + one bespoke line + filled fields |\n| Polish for an hour, send nothing | Misses the FRT window | Acknowledge fast, resolve honestly later |\n| One flat SLA for every ticket | Outage waits behind a typo | 3–4 priority tiers with distinct FRT targets |\n| Argue with an angry customer | Volume escalates, churn rises | Acknowledge feeling, own it, give a timed step |\n| Invent a policy \u002F price \u002F ETA | Hallucinated answers create real liability | Quote the KB or say \"let me confirm\" + escalate |\n| Escalate a bare ticket link | Next agent restarts, +90–180s, anger | Attach the handoff packet every time |\n| Leave a `{{placeholder}}` in the reply | Worse than no name; signals automation | Verify all fields filled before send |\n| \"Please calm down\" \u002F \"as I said\" | Tells them their feeling is the fault | De-escalation moves; banned-phrase swaps |\n| Promise a fix you can't verify | Breaks trust when it slips | Promise the *next update time*, not the fix |\n| Chase NPS off a single ticket | Wrong altitude metric | CSAT\u002FFCR per ticket; NPS at relationship level |\n","# Customer Support — Run the Desk on a Live Ticket\n\nYou own the **support desk as an operation**: take an incoming ticket, assign priority and intent, start the SLA clock, draft the reply (macro or bespoke), enforce tone, and decide escalation with a clean handoff. The output is a triaged, answered ticket — human judgment, not a code artifact.\n\n## The triage-to-reply loop (the spine)\n\nRun every ticket through these six steps in order. Skipping triage to \"just reply fast\" is how you send the wrong tone to a P1.\n\n```text\n1 Classify intent   →  what does the customer actually want? (bug, how-to, refund, outage, rage)\n2 Set priority      →  P1–P4 (impact × urgency × account tier)\n3 Start SLA clock   →  pick the FRT target NOW, before drafting\n4 Draft reply       →  macro skeleton or bespoke; one personalized line minimum\n5 Tone \u002F voice check →  on-brand, de-escalated, no banned phrases\n6 Escalate or send  →  trigger met? hand off with the packet. else send + log.\n```\n\n## Priority, SLA, and the escalation call\n\nThe branch is at step 6. This table sets priority, both SLA clocks, and whether the ticket leaves your hands:\n\n| Signal in the ticket | Priority | First response (FRT) | Resolution target | Escalate? |\n|---|---|---|---|---|\n| Production down \u002F data loss \u002F payments broken \u002F many users blocked | P1 critical | 15–30 min | ASAP, status updates every 30–60 min | Yes → tier-2\u002Feng + the packet |\n| Enterprise\u002FVIP account, or \"cancel \u002F refund \u002F lawyer \u002F unacceptable\" | P1–P2 | 30 min – 1 h | Same business day | Yes → owner + tier-2 |\n| One user blocked, no workaround | P2 high | 1–2 h | Same business day | Only if unresolved past target |\n| Question with a workaround, billing query | P3 medium | 4–8 h | 1–2 business days | No, unless policy claim is unverifiable |\n| FAQ, cosmetic, feature request | P4 low | ~1 business day | Best effort | No |\n\n**Why pick the FRT target before drafting:** the SLA decides your tone and length. A P1 gets a two-line acknowledgment in 20 minutes, not a polished essay in two hours.\n\nUse **3–4 priority tiers, never a flat SLA.** Too many tiers confuse the desk; too few don't differentiate the outage from the typo (Emailmeter \u002F Hiver, 2026-06-02).\n\n**Acknowledge fast even when you cannot resolve fast.** First-response and resolution are *separate, linked* SLAs: FRT sets the expectation, resolution is the outcome. 89% of customers feel valued by a fast first response even when the full fix takes longer (getMonetizely \u002F Hiver, 2026-06-02). So: acknowledge inside the FRT window with an honest \"here's what I know and when I'll update you,\" then resolve honestly. Never go silent to wait for a perfect answer.\n\nFull P1–P4 matrix by channel (email\u002Fchat\u002Fphone) lives in [`references\u002Fmacros-and-sla.md`](references\u002Fmacros-and-sla.md).\n\n## Macros done right\n\nA macro with **one genuinely personalized sentence raises CSAT; a verbatim template lowers it.** Macros cut first-response time 30–50%, but sent raw they read as scripted (TextExpander \u002F Gorgias, 2026-06-02). So every macro is a skeleton you finish, not a message you forward.\n\nThe 4-beat skeleton — every reply hits these in order:\n\n```text\n1 Empathy \u002F ack   →  name the problem in their words (1 line)\n2 Answer          →  the fix, the status, or the next concrete step\n3 Next step       →  what happens now + when they'll hear back\n4 Close           →  human sign-off, door left open\n```\n\nRules:\n\n- **Fill dynamic fields, then add one bespoke line.** Placeholders like `{{first_name}}`, `{{order_id}}`, `{{ticket_id}}`, `{{date}}` are table stakes — they are not personalization. The personalized line references *this* ticket's specifics.\n- **Never send a macro with a visible `{{placeholder}}`.** A leaked `{{first_name}}` is worse than no name.\n- **One reply, one ask.** If you need three things from the customer, you'll get one. Batch the request or sequence it.\n\nBad → Good (info-request macro):\n\n```text\nBad (verbatim template, robotic):\n  Dear Customer, Thank you for contacting support. We have received your\n  request. In order to assist you, please provide additional information.\n  Your satisfaction is important to us. Regards, The Support Team.\n\nGood (skeleton + one personalized line):\n  Hi Marta — sorry the export keeps failing on you. To dig in I need one\n  thing: the rough time you last hit \"Export CSV\" today (so I can pull the\n  matching server log). I noticed your account is on the new billing plan,\n  so this might be the same migration glitch we patched yesterday — I'll\n  confirm and get back to you within 2 hours. — Eric\n```\n\nThe Good version names the feature, gives a real next step + a time, and the \"new billing plan \u002F migration glitch\" line proves a human read the ticket.\n\nFull skeleton library (ack, info-request, resolved, refund, known-bug, outage, churn-save, escalation-handoff, feature-request decline, gentle-close) is in [`references\u002Fmacros-and-sla.md`](references\u002Fmacros-and-sla.md).\n\n## Tone and de-escalation\n\nApply the brand voice; do not invent it here. The voice guide is authored by [`..\u002Fbrand-voice\u002FSKILL.md`](..\u002Fbrand-voice\u002FSKILL.md) — this skill reads it and writes the reply against it.\n\nFor an angry or churn-risk ticket, de-escalation moves in order:\n\n1. **Acknowledge the feeling before the facts.** \"Three days without a reply is not okay, and I get why you're frustrated\" lands before any fix.\n2. **Take ownership in the first person.** \"I dropped this\" beats \"the team was unable to.\" No passive voice to hide behind.\n3. **One concrete next step with a time.** Vague reassurance reads as a brush-off; \"I'll have an answer by 3pm\" is trust.\n4. **Match their stakes, not their volume.** Stay calm and specific. Do not argue, and never match anger with anger.\n\nBanned phrases — they escalate, not de-escalate:\n\n| Never write | Why it backfires | Write instead |\n|---|---|---|\n| \"Please calm down.\" | Tells them their feeling is the problem. | \"You're right to be frustrated — let's fix it.\" |\n| \"As I said \u002F as previously mentioned\" | Scolds them for not reading. | Just restate it plainly. |\n| \"Per our policy.\" (naked) | Hides behind rules, no human. | Explain the *why*, then the policy. |\n| \"Unfortunately, there's nothing we can do.\" | Dead end, zero path. | Offer the nearest thing you *can* do. |\n| \"This is a known issue.\" (alone) | Sounds like \"we knew and didn't care.\" | Add the status + the ETA or the workaround. |\n\nChecklist before send: feeling acknowledged · first-person ownership · one timed next step · no banned phrase · no `{{placeholder}}` left.\n\nDe-escalating *this* ticket is the job here; running a win-back or renewal **program** is `retention`.\n\n## Escalation\n\nEscalate on **explicit conditions, not vibes.** Standard tiers: T1 frontline\u002Fcommon, T2 specialist\u002Fcomplex, T3 engineering or exec\u002Fcritical (Kapture \u002F Hiver, 2026-06-02). Hand off when any trigger fires:\n\n- **Negative-language signals:** \"unacceptable\", \"cancel\", \"refund\", \"lawyer\", \"chargeback\".\n- **SLA at risk:** the FRT or resolution target is about to blow.\n- **Account tier:** enterprise \u002F VIP.\n- **Confidence below threshold:** if an assisted\u002FAI draft is below ~85% confidence on a policy or price claim, hand to a human — don't guess (swiftflutter \u002F Kapture, 2026-06-02).\n\n**Context must travel with the escalation.** 70% of customers expect the next agent to know their history; making them repeat themselves adds 90–180 seconds per ticket (Forrester via Fini Labs, 2026-06-02). Never escalate a bare ticket link. Attach the handoff packet:\n\n```text\nHANDOFF PACKET\n- Customer + account tier:  Marta R. — Enterprise (VIP)\n- Ticket \u002F priority:        #4821 — P1\n- One-line summary:         CSV export 500s since ~09:00; blocking month-end close.\n- Steps already tried:      Confirmed not browser\u002Fcache; reproduced on staging.\n- Customer sentiment:       Angry, has mentioned \"cancel\" once.\n- Collected variables:      plan=new-billing, last_export=09:14, region=eu-west\n- What I need from you:      Confirm if the billing-migration patch covers this.\n```\n\nThe full escalation matrix (trigger → tier → owner) is in [`references\u002Fmacros-and-sla.md`](references\u002Fmacros-and-sla.md).\n\n## FAQ and KB answering\n\n**Quote the KB or escalate — never invent policy, price, or an ETA.** Hallucinated support answers produced 150+ documented legal cases by mid-2025 (cmswire \u002F swiftflutter, 2026-06-02). The guardrail:\n\n- If the answer is in the KB, **quote it and link the article.** Don't paraphrase a refund window or a price from memory.\n- If it is not in the KB, say \"let me confirm\" and escalate — do not fill the gap with a plausible guess.\n- Deflection is **bounded by KB quality.** Triage-only without a strong KB deflects ~20–30%; well-organized KBs push assisted\u002Fauto resolution to 50–80% (Zendesk \u002F Intercom \u002F Fini Labs, 2026-06-02). A bad answer fast is worse than a correct answer slightly slower.\n\nThis skill *consumes* the KB; it never writes it. If the article the customer needs does not exist, flag the gap — authoring it is a `technical-writing` job, and building the autonomous auto-answer bot on top of the KB is `chatbot` + `rag`.\n\n## Metrics — what to watch and when\n\nNo single metric is sufficient; track the trio plus resolution rate (Armatis \u002F Giva, 2026-06-02):\n\n| Metric | What it measures | When to optimize for it |\n|---|---|---|\n| **CSAT** | Transactional satisfaction; `positive ÷ total × 100` | After any single ticket; the default health signal |\n| **CES** | Customer effort, 1–7 scale | When tickets resolve but customers still churn — friction is the problem |\n| **NPS** | Relationship; promoters − detractors | Quarterly \u002F relationship level, not per-ticket |\n| **FCR \u002F one-touch** | Resolved in one interaction | When repeat contacts are climbing — optimize the first reply |\n\nRule of thumb: optimize **FRT** for reassurance (the customer feeling handled), **FCR** for effort (fewer round-trips). They pull in different directions — don't chase both blindly.\n\n## Anti-patterns\n\n| Anti-pattern | Why it fails | Do instead |\n|---|---|---|\n| Send the macro verbatim | Reads as scripted, drops CSAT | Skeleton + one bespoke line + filled fields |\n| Polish for an hour, send nothing | Misses the FRT window | Acknowledge fast, resolve honestly later |\n| One flat SLA for every ticket | Outage waits behind a typo | 3–4 priority tiers with distinct FRT targets |\n| Argue with an angry customer | Volume escalates, churn rises | Acknowledge feeling, own it, give a timed step |\n| Invent a policy \u002F price \u002F ETA | Hallucinated answers create real liability | Quote the KB or say \"let me confirm\" + escalate |\n| Escalate a bare ticket link | Next agent restarts, +90–180s, anger | Attach the handoff packet every time |\n| Leave a `{{placeholder}}` in the reply | Worse than no name; signals automation | Verify all fields filled before send |\n| \"Please calm down\" \u002F \"as I said\" | Tells them their feeling is the fault | De-escalation moves; banned-phrase swaps |\n| Promise a fix you can't verify | Breaks trust when it slips | Promise the *next update time*, not the fix |\n| Chase NPS off a single ticket | Wrong altitude metric | CSAT\u002FFCR per ticket; NPS at relationship level |\n",{"name":5,"tags":24,"origin":32,"recommends":33,"description":7},[5,25,26,27,28,29,30,31],"helpdesk","ticket-triage","sla","escalation","macros","de-escalation","csat","risco",[34,35,36,37,38,39],"technical-writing","client-onboarding","retention","chatbot","rag","brand-voice","main","d69096e239f2a89aabc2cd078c47664a9d3b1e11","03-逆向工程\u002Fcustomer-support",3,"https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts","https:\u002F\u002Fgithub.com\u002Flza6","个人vx：Tf00798\r\n个人✈：@TFGY999","2026-10-08 14:29:22",{"source":8,"skillName":5,"commands":49,"prompt":59,"zipAvailable":60},[50,53,56],{"tool":51,"command":52},"npx","npx skills add lza6\u002FArmor-piercing-system-prompts --skill customer-support",{"tool":54,"command":55},"bunx","bunx skills add lza6\u002FArmor-piercing-system-prompts --skill customer-support",{"tool":57,"command":58},"pnpm","pnpm dlx skills add lza6\u002FArmor-piercing-system-prompts --skill customer-support","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F03-逆向工程\u002Fcustomer-support\u002FSKILL.md",true,[62,68,72],{"path":63,"resourceType":64,"resourceTypeDesc":65,"sizeBytes":66,"sourceHtmlUrl":67},"evals\u002Fcases.yaml","REFERENCE","参考文档",4353,"https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F03-逆向工程\u002Fcustomer-support\u002Fevals\u002Fcases.yaml",{"path":69,"resourceType":64,"resourceTypeDesc":65,"sizeBytes":70,"sourceHtmlUrl":71},"evals\u002FREADME.md",836,"https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F03-逆向工程\u002Fcustomer-support\u002Fevals\u002FREADME.md",{"path":73,"resourceType":64,"resourceTypeDesc":65,"sizeBytes":74,"sourceHtmlUrl":75},"references\u002Fmacros-and-sla.md",5651,"https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F03-逆向工程\u002Fcustomer-support\u002Freferences\u002Fmacros-and-sla.md",[77,85,90,97,104,111],{"skillNo":78,"name":79,"nameZh":6,"description":80,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":81,"categories":82,"keywords":83,"publishedAt":84},"01M7FFG485KY83TD6HSC4YPCMC","obsidian-clipper-template-creator","Guide for creating templates for the Obsidian Web Clipper. Use when you want to create a new clipping template, understand available variables, or format clipped content.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobsidian-clipper-template-creator\u002FSKILL.md",[],[],"2026-11-16 10:05:34",{"skillNo":86,"name":79,"nameZh":6,"description":80,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":87,"categories":88,"keywords":89,"publishedAt":84},"01M7FFG0T6AZ0STEJPSB3CX84K","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobsidian-clipper-template-creator-antigravity-awesome-skills-main\u002FSKILL.md",[],[],{"skillNo":91,"name":92,"nameZh":6,"description":93,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":94,"categories":95,"keywords":96,"publishedAt":84},"01M7FFFY0QEM14SKWPDYGMXM2P","python-api","Programmatic Object-Detection-Metrics API guidance for BoundingBox, BoundingBoxes, Evaluator, enums, coordinate conversion, AP\u002FmAP semantics, and source-style import diagnostics.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobject-detection-metrics\u002Fsub-skills\u002Fpython-api\u002FSKILL.md",[],[],{"skillNo":98,"name":99,"nameZh":6,"description":100,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":101,"categories":102,"keywords":103,"publishedAt":84},"01M7FFFVZQZ2GJ0C0TGP7JHD20","file-evaluation","Folder-based PASCAL VOC AP\u002FmAP evaluation from ground-truth and detection text files.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobject-detection-metrics\u002Fsub-skills\u002Ffile-evaluation\u002FSKILL.md",[],[],{"skillNo":105,"name":106,"nameZh":6,"description":107,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":108,"categories":109,"keywords":110,"publishedAt":84},"01M7FFFSFNYSE77SJ6PE1CK2DA","object-detection-metrics","Guides legacy Object-Detection-Metrics PASCAL VOC AP\u002FmAP evaluation from detection text folders and direct Python API classes.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobject-detection-metrics\u002FSKILL.md",[],[],{"skillNo":112,"name":113,"nameZh":6,"description":114,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":115,"categories":116,"keywords":117,"publishedAt":84},"01M7FFFP0GCK1RVXBCPQ8APBMT","nx-workspace-patterns","Configure and optimize Nx monorepo workspaces. Use when setting up Nx, configuring project boundaries, optimizing build caching, or implementing affected commands.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fnx-workspace-patterns\u002FSKILL.md",[],[]]