health-score
SolidUse when reviewing customer-account health with a configured scoring rubric and dated inputs, or when reporting why a portfolio cannot yet be scored. Not for finding expansion opportunities; use expansion-opportunities.
Install
Quality Score: 83/100
Skill Content
Details
- Author
- davekilleen
- Repository
- davekilleen/Dex
- Created
- 7 months ago
- Last Updated
- yesterday
- Language
- Python
- License
- NOASSERTION
Integrates with
Bundled in these plugins
Similar Skills
Semantically similar based on skill content — not just same category
health-score-designer
When the user wants to design, weight, calibrate, backtest, audit, or fix a customer health score. Also use when the user mentions 'health score is garbage', 'nobody trusts it', 'score that actually predicts', 'design a health score', 'build a customer health score', 'our health score is broken', 'nobody trusts the health score', 'what should go into our health score', 'health score weights', 'green accounts keep churning', 'is our health score any good', 'backtest the health score', 'calibrate our health score', 'red yellow green thresholds', 'the score isn't predicting anything', 'audit our scoring model', or 'rescore the book'. Use this whenever someone is deciding what makes an account healthy, or arguing about a score's inputs, even if they never say 'health score' — a request for 'one number that tells me who to work on' is this skill. For scoring one account today, see churn-risk. For the data inventory the score is built on, see cs-context. For per-account renewal probability, see renewal-forecast.
account-health-scorer
Score and rank an entire portfolio of enterprise accounts for churn risk and expansion opportunity from a single portfolio export. Use this skill whenever the user mentions portfolio review, renewal risk, churn triage, book-of-business health, account prioritization, or uploads a multi-account CSV and asks which accounts need attention, even if they do not say "score."
health-read
Audits an account health score rather than reporting it, so the number can be trusted or discounted deliberately. Trigger whenever the user says "health score", "account health", "is this account healthy", "the score says green but", "health check", "why is this account red", "our health scores are wrong", "audit the health score", or asks what an account's health actually is. Also trigger when a green account has just churned, or when someone is deciding whether to act on a score. It separates what is measured from what is inferred, finds the inputs that are proxies wearing the clothes of evidence, checks whether the composite is averaging a growing signal against a dying one, and asks the question nobody asks: has this score ever been tested against accounts that actually left. Runs on whatever the user can describe about their scoring, and says plainly which parts of the read are unsupported.