Agent Skills
› mohitagw15856/pm-claude-skills
› cs-health-scorecard
cs-health-scorecard
GitHub为客户账户构建结构化的健康评分卡，评估续保风险与扩展潜力。基于产品采用、参与度、成果等维度打分，生成RAG状态及关键风险和建议行动，支持通过Brain读写数据并调用Python脚本自动计算得分。
Trigger Scenarios
评估账户健康状况
判断续保风险
构建健康仪表盘
评估账户续约或扩展可能性
Install
npx skills add mohitagw15856/pm-claude-skills --skill cs-health-scorecard -g -y
SKILL.md
Frontmatter
{
"name": "cs-health-scorecard",
"description": "Build a customer health scorecard for a specific account. Use when asked to score account health, assess renewal risk, build a health dashboard, or evaluate an account's likelihood to renew or expand. Produces a structured health scorecard with a RAG status, dimension scores, key risks, and recommended actions."
}
Customer Health Scorecard Skill
Produce a structured, data-driven health scorecard for a customer account — giving the CSM and leadership a clear view of renewal risk, expansion potential, and the actions needed to move the account in the right direction.
Reads from / Writes to the Brain
If a professional-brain (brain/) exists, ground in it instead of re-asking for what you already know:
- Read first: the account's
entities/file, itsstakeholders/(champion, economic buyer, detractors), andknowledge/. Runpython3 ../professional-brain/scripts/brain_query.py ./brain "<account name>"and carry each fact's provenance tag through. - 📥 Propose to the Brain: after producing, propose recording the health verdict + key risks to the account
entities/file, and a renewal-risk entry todecisions/if a call is made, each provenance-tagged. Show them, get a yes, then write with../professional-brain/scripts/brain_write.py … --commit(append-only, dry-run by default).
Required Inputs
Ask for these if not already provided:
- Account name and tier (enterprise / mid-market / SMB)
- Contract value (ARR) and renewal date
- Product usage data — logins, DAU/MAU ratio, key feature adoption
- Support data — open tickets, CSAT or NPS score, recent escalations
- Engagement data — last QBR date, executive sponsor status, champion name
- Commercial data — payment history, expansion conversations, seats used vs. licensed
- Any known risks or recent changes at the account
Scoring Framework
Score each dimension 1–5. Weight as shown. Calculate weighted total out of 100.
| Dimension | Weight | What to Score |
|---|---|---|
| Product Adoption | 30% | DAU/MAU ratio, breadth of features used, power users identified |
| Engagement | 20% | QBR cadence, executive sponsor active, champion strength |
| Outcomes | 20% | Customer hitting their stated goals / success metrics |
| Support Health | 15% | Ticket volume trend, unresolved escalations, CSAT |
| Commercial | 15% | On-time payments, seats utilised, expansion signals |
Score → RAG conversion:
- 80–100: Green (healthy, renew likely)
- 60–79: Amber (at risk, needs attention)
- 0–59: Red (high churn risk, escalate)
Programmatic Helper
This skill ships with a stdlib-only Python script that applies the weights above and converts the weighted total to a RAG status — so the headline score is computed identically every time and weights always sum to 100%.
# Five scores 1-5 in order: adoption engagement outcomes support commercial
python3 scripts/health_score.py --scores 4 3 4 2 5 --account "Acme Corp"
# Or from JSON (lets you override the default weights per account/segment)
python3 scripts/health_score.py --input account.json
It returns the per-dimension weighted points, the total out of 100, and the RAG band (Green ≥80, Amber 60–79, Red <60) with a one-line next step. Run it to set the headline number, then write the dimension detail and actions below around it. Add --json for downstream tooling.
Output Format
Customer Health Scorecard: [Account Name]
CSM: [Name] | Tier: [Enterprise / Mid-Market / SMB] ARR: £/$/€[X] | Renewal date: [Date] | Days to renewal: [N] Overall health: [Green / Amber / Red] — [Score]/100 Last updated: [Date]
Health Score Summary
| Dimension | Score (1–5) | Weight | Weighted Score | Trend |
|---|---|---|---|---|
| Product Adoption | [1–5] | 30% | [X] | ↑ / → / ↓ |
| Engagement | [1–5] | 20% | [X] | ↑ / → / ↓ |
| Outcomes | [1–5] | 20% | [X] | ↑ / → / ↓ |
| Support Health | [1–5] | 15% | [X] | ↑ / → / ↓ |
| Commercial | [1–5] | 15% | [X] | ↑ / → / ↓ |
| Total | — | 100% | [X]/100 |
Dimension Detail
Product Adoption — [Score]/5
- DAU/MAU ratio: [X]% (benchmark: >25% = healthy)
- Key features adopted: [List features in use]
- Features not adopted: [List unused high-value features]
- Power users identified: [Yes / No — how many]
- Assessment: [1–2 sentences on adoption health]
Engagement — [Score]/5
- Last QBR: [Date] — [Outcome summary]
- Next QBR: [Scheduled / Overdue]
- Executive sponsor: [Active / Passive / Vacant]
- Champion: [Name, role, strength: strong / moderate / weak]
- Assessment: [1–2 sentences]
Outcomes — [Score]/5
- Customer's stated goals: [List 2–3 goals from onboarding or last QBR]
- Progress against goals: [On track / Partial / Off track]
- Evidence of value: [Metric or quote that demonstrates ROI]
- Assessment: [1–2 sentences]
Support Health — [Score]/5
- Open tickets: [N] (priority breakdown: P1: X, P2: X, P3: X)
- CSAT / NPS: [Score] (benchmark: >8 CSAT / >30 NPS = healthy)
- Unresolved escalations: [Yes / No — details if yes]
- Ticket trend (last 90 days): Increasing / Stable / Decreasing
- Assessment: [1–2 sentences]
Commercial — [Score]/5
- Seats licensed: [N] | Seats active: [N] ([X]% utilisation)
- Payment history: [On time / Late — details]
- Expansion signals: [Yes — describe / No]
- Downgrade or cancellation signals: [Yes — describe / No]
- Assessment: [1–2 sentences]
Top Risks
| Risk | Severity | Mitigation |
|---|---|---|
| [Risk description] | High / Medium / Low | [Specific action to mitigate] |
Recommended Actions
Immediate (this week):
- [Action — owner — deadline]
This month:
- [Action — owner — deadline]
Before renewal:
- [Action — owner — deadline]
Renewal Forecast
| Scenario | Probability | ARR at risk |
|---|---|---|
| Full renewal at current ARR | [X]% | £/$/€0 |
| Renewal with contraction | [X]% | £/$/€[X] |
| Churn | [X]% | £/$/€[full ARR] |
Recommended renewal play: [Expand / Hold / Save / Manage out]
Deeper Materials
This skill ships with support files — use them when they are available:
references/leading-signals.md— Health Signals That Lead (Instead of Eulogise). Apply it while producing the output; it carries the calibration and judgment calls the method summary above compresses.templates/account-scorecard.md— a fill-in version of the deliverable with the quality gates inline. Offer it when the user wants to work the document themselves rather than have it generated.
Scoring Rubric (0–40)
Score any output of this skill before handing it over; 32+ is ship-quality.
| Dimension | 0 | 5 | 10 |
|---|---|---|---|
| Score integrity | Weighted total doesn't compute from the dimension scores and stated weights, or the RAG band contradicts the total | Arithmetic is correct but weights were adjusted silently, or the headline RAG smooths over a dimension that tells a different story | Total computes exactly (score × weight on the 1–5 scale, out of 100), RAG matches the 80/60 bands, and any dimension that contradicts the overall status is called out rather than averaged away |
| Evidence per dimension | Dimension scores asserted with no supporting data ("engagement feels weak") | Most dimensions cite data, but at least one score leans on gut feel or a stale data point presented as current | Every dimension score is anchored to named, dated evidence (usage figures, ticket counts, QBR dates, seat utilisation) and benchmarks are applied where the format provides them |
| Risk specificity | Risks are labels ("low engagement", "churn risk") with no people, dates, or dollar amounts | Risks are real but partially vague — severity assigned without a mitigation, or mitigations without owners | Every risk names the person/event/amount involved ("champion departs 25 July, no successor"), carries a severity, and has a mitigation someone could start this week |
| Renewal calibration | Forecast missing, probabilities don't sum to 100%, or the recommended play ignores the score | Forecast present and sums correctly, but ARR-at-risk figures don't reconcile to contract line items, or the play is generic | Probabilities sum to 100%, ARR at risk maps to actual contract components, the play (Expand/Hold/Save/Manage out) follows from the score and risks, and actions are owned, dated, and sequenced against the renewal date |
Quality Checks
- Score is based on data, not gut feel — each dimension has evidence
- Risks are specific (not "low engagement" — something like "executive sponsor left in March, no replacement identified")
- Actions have owners and deadlines
- Renewal probability is calibrated against pipeline reality
- Trend arrows reflect direction of change vs. last scorecard, not just current state
Anti-Patterns
- Do not score health dimensions on gut feel — every score needs specific supporting evidence
- Do not give a Green status to accounts with unresolved P1 issues or missed milestones
- Do not list risks vaguely — "low engagement" without specifics is not actionable
- Do not leave recommended actions without named owners and deadlines
- Do not conflate product usage frequency with product value delivery
Version History
- 54fad50 Current 2026-07-19 12:50
- a38bc30 2026-07-05 11:14
