Plugin health
Plugin health says, per plugin version, how many calls the instances that opted in to community telemetry reported, and how many worked. A version is deemed validated with at least 200 reports and 90 % of them successful, over the last 90 days; below the minimum of reports the site says "not enough reports yet" and shows no rate.
The collector could not be read when this site was built, so no figure is shown.
| Plugin | Version | Health |
|---|---|---|
| Cloudflare DNS | 3.0.0 | not enough reports yet |
| Context7 | 0.1.0 | not enough reports yet |
| Fly.io | 0.1.0 | not enough reports yet |
| GitHub | 3.2.0 | not enough reports yet |
| Gmail | 0.3.0 | not enough reports yet |
| Miro | 0.1.0 | not enough reports yet |
| Notion | 0.1.0 | not enough reports yet |
| Railway | 4.2.0 | not enough reports yet |
| Resend | 0.3.0 | not enough reports yet |
| SSH (bash) | 2.0.0 | not enough reports yet |
| Stripe | 0.4.0 | not enough reports yet |
| Supabase | 4.2.0 | not enough reports yet |
| Upstash Redis | 3.0.0 | not enough reports yet |
| Vercel | 3.0.0 | not enough reports yet |
How it is measured
- Telemetry is off on every instance unless its operator turns it on; what an instance sends is listed on the trust center.
- Every outcome that is not a success counts against the rate, including a provider refusing a call the agent got wrong: the rate measures calls that worked, not the plugin's code alone.
- Anyone can send reports, so a determined party could skew a plugin's figures; the thresholds and a global ingestion limit raise the cost, they do not remove it. The status is a community signal, not a security guarantee.
- The thresholds live in one SQL function of the collector,
plugin_validation_thresholds(), published in the collector's README.