client-health-radar

Detects client churn risk and scope creep from meetings, threads, and captured dashboards.

84|37|Updated May 23, 2026
One-click install
npx skills add https://github.com/legioncodeinc/vibe-coding-tools --skill client-health-radar-legioncodeinc
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: client-health-radar
Source: https://github.com/legioncodeinc/vibe-coding-tools/tree/main/src/quarantine/skills/keep/client-health-radar
Command: npx skills add https://github.com/legioncodeinc/vibe-coding-tools --skill client-health-radar-legioncodeinc

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Service businesses lose clients they thought were fine and give away unbilled work they never tracked. This Skill reads your recorded meetings, message threads, and captured dashboards per client and surfaces churn signals and scope creep as dated, quoted evidence instead of a gut-feel score. ## Core Features & Use Cases - Five signal families per client: unmet promises in both directions, silence gaps against each client's own cadence baseline, scope creep with the exact quote where the ask happened, room composition changes, and payment or renewal signals. - Bands, never scores: clients are banded Green, Amber, Red, or Unknown with named evidence and receipts, deliberately refusing to emit a health score or sentiment score. - Weekly routine plus deep dives: sets up a Monday 07:30 routine that reports what changed, and runs 90-day deep dives on demand that write a ranked risk report file. - Use Case: A freelancer notices a client has gone quiet. Running the skill shows the client is 26 days past their normal 7-day contact gap, requested source-file exports two weeks ago, and has an approval open for 31 days, with one recommended action: call a named contact. ## Quick Start Ask your assistant to run a client health check across your roster to see which clients are at risk and which are accumulating out-of-scope work.

Frequently Asked Questions about client-health-radar

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I detect which client is about to churn?▼

Run a client health check that reads your recorded meetings, threads, and dashboards per client. It flags silence gaps against each client's own cadence baseline, unmet promises, exit-shaped requests like asset exports, and commercial signals, each with a dated receipt.

How to track scope creep on a client engagement?▼

The skill establishes a scope baseline from kickoff meetings or statements of work, then finds out-of-scope asks in meetings and threads. Every flagged item carries the exact quote with meeting name and date, split into absorbed, open, and handled buckets.

What tools does client health monitoring require?▼

It requires the Littlebird MCP connected on a Power or Pro plan, plus recorded meetings and a client roster you confirm once. Without Littlebird tools available in the session, the skill stops rather than guessing from other sources.

Why does the skill refuse to produce a client health score?▼

Research shows transcription substitutes the emotion-carrying word in roughly one utterance in six, and composite scores predict almost nothing. It reports four bands with quoted evidence instead, and treats Unknown as a real band when coverage is thin.

Can the skill send follow-up messages to at-risk clients?▼

No. It produces an internal view only and never sends anything. Drafted outreach is held for approval with the full text shown first, and the skill hands approved text back to you rather than sending it.

What happens when no data is found for a client?▼

The client is banded Unknown, not Green, with the queries run and coverage counts reported. The recommended action is to get the next conversation recorded, since a client the skill cannot see is a coverage problem.