Tools
Small tools that run in this page: check an evidence bundle, read a security.txt, defang URLs. Each loads what it needs the first time you use it, and nothing you pick or paste leaves the page. Also: unwrap a URL, and check the AI agents your log declares against their providers' published ranges and keys.
Check an evidence bundle
Got a report? Pick the files attached to the email.
Pick the files of a bundle from swarmchase case bundle (records.ndjson, analysis.json, stamp.json, MANIFEST.sha256, and MANIFEST.sha256.sig when it is signed) or its folder. The page checks the manifest's signature against the key this site publishes, each SHA-256 against the manifest, re-runs the analysis with the bundled rules (or your edited ones), and compares every tier and cluster with the stamp, as swarmchase case verify does. It then shows the method and evaluation the stamp cites, the attribution layers the records support (add a claims.json from case work to include its claims), the runs, any candidate fleets and escalations, and the records. The files are read in this page only.
No bundle to hand? The showcase bundle is synthetic: built with this site from the showcase records the home page walks through. None of it was collected.
Read a security.txt
Paste an organisation's /.well-known/security.txt to pull out who to notify. Or run swarmchase contact <domain>.
Defang
Make URLs safe to paste into reports and chat: hxxps://example[.]org/. Every http and https URL in the text changes, nested ones too: its scheme, and the dots of its host. Bare host names and IP addresses outside a URL are left as typed.