Every agent's model. One place. Codex on DeepSeek, Claude Code on Kimi, from the menu bar.
masking: comments no longer quote Routing's "masking…" (#4053)
yetone174250b277bb
diff --git a/internal/masking/masking.go b/internal/masking/masking.goindex fb5e20731..996cd0c98 100644--- a/internal/masking/masking.go+++ b/internal/masking/masking.go@@ -145,7 +145,7 @@ func toGuard() {// Past it the rules mask alone. AstrLink Guard reads on the CPU, a long// history it hasn't seen takes minutes there, and one request at a time// is read: unbounded, every request queued behind the first (shensmart-// on Discord, 10-28: Codex's requests sat "masking…" once the default's+// on Discord, 10-28: Codex's requests sat masking once the default's// download finished). What the model read meanwhile is kept, so the next// request finds it. A var for tests.var defaultWait = 2 * time.Seconddiff --git a/internal/masking/masking_test.go b/internal/masking/masking_test.goindex 82359e5af..9d8702321 100644--- a/internal/masking/masking_test.go+++ b/internal/masking/masking_test.go@@ -167,7 +167,7 @@ func (h hangs) Find(context.Context, []string) ([][]redact.Found, error) {}// shensmart on Discord (10-28): with the model on by default, a request-// waited for it as long as its client did, "masking…" on the Requests+// waited for it as long as its client did, masking on the Requests// page, every one after it queued behind. On by default, it is waited for// defaultWait at most and the rules mask alone; turned on by the user, it// is waited for longer (strictWait).