PulseFeed is run by one operator; every message below lands in the same inbox and is read by a person. · Home · Observatory · Methodology
| You want to… | Best route |
|---|---|
| Report a security vulnerability in PulseFeed itself | security.txt (RFC 9116) → email with a clear subject. Please give us a chance to fix before public disclosure. |
| Dispute a verdict, a trust score or a number we publish | Email or a GitHub issue. Every published figure is recomputable from open data — tell us which number and we will recount it in public; corrections get their own page. |
| Ask your endpoint or MCP package to be re-probed | GitHub issue with the URL or package name. Probes are read-only and free. |
| Use the datasets, join series, or collaborate on research | Email. Frozen dataset with recount scripts: doi:10.5281/zenodo.22270727 (CC BY 4.0); daily manifests are signed and Bitcoin-timestamped (/anchors/latest.json). |
| Get webhook alerts for your own packages | No email needed: POST /mcp/alerts/subscribe — see alerts status. |
Replies usually within a day. If a figure of ours is shown wrong, the correction is published, dated and kept — see the corrections page for how that has worked in practice. We do not send marketing mail and we do not want any.
PulseFeed — independent trust & safety layer for the x402 agent economy. Operator: Nikolai Otrishko (independent researcher). Machine-readable security contact: /.well-known/security.txt.