{"version":"2.1.238","anchor":"safety-classifier-severity-scoring-extended-to-two-more-site","canonical_anchor":"safety-classifier-severity-scoring-extended-to-two-more-site","heading":"Safety-classifier severity scoring extended to two more sites, both inert here","tier":"soon","area":"Permissions","scope":null,"heads_up":null,"url":"https:\/\/changelogs.core-directive.com\/v\/2.1.238\/e\/safety-classifier-severity-scoring-extended-to-two-more-site","release_url":"https:\/\/changelogs.core-directive.com\/v\/2.1.238","markdown":"### Safety-classifier severity scoring extended to two more sites, both inert here\n\nThe safety classifier can score severity at agent handoff and sandboxed network access, but both stay silent.\n\n**What**\n\nThe safety classifier can now compute a severity score at two additional decision points, agent handoff and sandboxed network access, alongside the existing workflow gate. Each site carries per-model score thresholds. Severity is only computed when server-side config lists the site as on, so with no server value both new sites return no severity and change nothing.\n\n**Details**\n\n- Sites are named `handoff` and `sandboxNetwork`; the server key controlling them is `severityBySite`, and a missing or \"off\" entry yields a null severity.\n- Thresholds are a t1\/t2 pair per model name.\n- The default table names a model, `claude-opus-4-8`, at thresholds 45 and 35.\n\n**Evidence**\n\n`\"claude-opus-4-8\": { t1: 45, t2: 35 }`\n\n- Area: Permissions\n- Tier: Nothing to try yet\n- Useful: 2\/5\n- Signal: 4\/5\n- Present in the build but not switched on"}