Generates a comprehensive client health overview across all accounts. Reads CRM data, support tickets, usage metrics, billing, and engagement logs. Calculates health scores, trend direction, and RAG status per client. Outputs a sorted risk report with recommended actions.
SkillScore83Maintained
Automatic
◆Customer Success pack
Last verified 8/24/2026
What it changes
Passed 19 of 23 test cases with the skill, and 10 of 23 without it.
Without skill43%
With skill83%
Checks it passed that the model alone missed
Exclude emojis from the entire response.
Attribute each metric to its specific data source.
Include only BetaCorp and GammaInc in the report output.
and 7 more in the full report
9 fixed10 passed either way4 still failing0 broke
Pass-rate lift
+39 pp
Tokens
+18%
Agent turns
0%
Response time
-17%
Model gemini-3.6-flashMeasured 8/24/2026v119 of 23 cases passed with skill
Tokens, turns and time describe resource use, not a dollar price.
See it in action
One of 9 test cases the skill fixed. Same model, same prompt, with and without the skill.
1 of 9
Prompt
Before completing the client health document generation, what check must be performed regarding RAG assignments?
What was checked, without and with the skill
Checked
Without skill
With skill
Verify that RAG assignments correspond to calculated composite score ranges.
Missed
Passed
case-15 · gemini-3.6-flash · 8/24/2026 · answers are capped and redacted
What's inside
The instructions your model reads when the skill loads.
SKILL.md42 lines5 files
Client Health Dashboard
Generate a data-driven client health report: pull data from every available source, compute a weighted health score per client, and produce a prioritized risk report (client-health-report.md) sorted by risk with RAG status and actionable recommendations.
Contents
references/data-sources.md -- what to pull from CRM, support, usage, billing, and communication channels
references/output-format.md -- exact report structure, formatting rules, and missing-data handling
Workflow
Collect data from every available source. Handle failures gracefully: log what was unavailable and proceed with partial data. Never fabricate data. See references/data-sources.md for the full source list and the fields to extract per client.
Score each client. Rate the five dimensions 0-100, apply weights, and compute the composite score. Assign RAG status and trend direction. See references/scoring-model.md.
Analyze risk and generate recommendations. Flag critical and warning risk factors, produce 2-4 specific recommendations targeting each client's weakest dimensions, and assess expansion potential for healthy accounts. See references/risk-and-recommendations.md.
Generate the report. Write client-health-report.md following the exact structure and formatting rules. Handle missing data by scoring neutral (50) and noting gaps. See references/output-format.md.
Validate before finalizing:
Verify RAG assignments match score ranges.
Confirm section ordering and within-section sorting.
Confirm every client appears exactly once.
Confirm each client has 2-4 specific, actionable recommendations.
Attribute each data point to its source.
Mark data gaps explicitly; never invent data that was not retrieved.
Interaction
If the user specifies particular clients, filter the report to those only.
If the user specifies a data source, prioritize it.
If the user provides CSV/Excel files, parse them as a primary source.
If the user requests a format variation, adapt accordingly.
Confirm the output path before writing.
If no data sources are accessible, explain what is needed and what to provide.
Constraints
Never fabricate or hallucinate data; report only what was retrieved, attributed to its source.
Never include credentials, API keys, or PII beyond business contact info.
Keep health scores mathematically correct per the weighting formula.
Keep recommendations specific and actionable, not generic.