| name | monitor |
| description | KPI monitoring and issue proposal: analyze crash rates, store reviews, and metrics to suggest next priorities |
| user-invocable | true |
| disable-model-invocation | true |
| allowed-tools | ["Read","Grep","Glob","Agent","Bash(gh issue create:*)","Bash(gh issue list:*)","TaskCreate","TaskUpdate","TaskList","AskUserQuestion","ToolSearch"] |
/monitor — KPI Monitoring & Prioritization (PoC)
Status: Proof of Concept — Requires Firebase MCP, Store API access, and analytics integration.
This skill defines the target workflow for fully autonomous AI-driven operations.
Monitor project KPIs and propose next actions based on data.
Steps
1. Crashlytics Analysis (if Firebase configured)
- Use Firebase MCP to fetch recent crash data
- Identify top crashes by frequency and user impact
- For critical crashes (crash-free rate < 99.5%), create GitHub Issues automatically
2. Store Review Analysis (if applicable)
- Collect recent App Store / Google Play reviews
- Perform sentiment analysis: positive / neutral / negative
- Identify recurring complaints or feature requests
- For strong negative trends, create GitHub Issues with
user-feedback label
3. Metrics Dashboard
Collect and report on key metrics:
| Metric | Source | Threshold |
|---|
| Crash-free rate | Crashlytics | < 99.5% → Issue |
| D1 Retention | Analytics | < 40% → investigate |
| D7 Retention | Analytics | < 20% → investigate |
| D30 Retention | Analytics | < 10% → investigate |
| Feature usage rate | Analytics | < 5% → consider removal |
| Conversion rate | Store | declining → investigate |
4. Feature Prioritization (2-Axis)
Score potential features on two axes:
Axis 1: User Requests (Qualitative)
- Request volume (votes, mentions, reviews)
- Sentiment intensity
- User segment (free/paid, engagement level)
Axis 2: Metrics (Quantitative)
- Impact on retention
- Impact on conversion
- Impact on crash-free rate
- Development effort (S/M/L)
Rule: Only features where both axes align get proposed. Exception: crash/security fixes act on metrics alone.
5. Issue Proposal
For each recommended action:
gh issue create \
--title "{priority-label}: {concise description}" \
--body "{rationale with data points}" \
--label "ai-proposed,{priority}"
Output Format
## KPI Monitor Report
**Date:** YYYY-MM-DD
**Period:** last 7 days
### Health
| Metric | Current | Trend | Status |
|---|---|---|---|
### Issues Created
| # | Title | Priority | Rationale |
|---|---|---|---|
### Recommended Next Actions
1. ...
2. ...