reduce seed iterations to 150M (~5-10s on modern hardware)

1B iterations was too slow (30s+). Benchmarked on Apple Silicon:
- 150M iterations ≈ 6.3s
- Falls within the 5-10s target range

internal/cli/mfer.go

@@ -156,7 +156,7 @@ func (mfa *CLIApp) run(args []string) {
 					},
 					&cli.StringFlag{
 						Name:    "seed",
-						Usage:   "Seed value for deterministic manifest UUID (hashed 1B times with SHA-256)",
+						Usage:   "Seed value for deterministic manifest UUID (hashed 150M times with SHA-256, ~5-10s)",
 						EnvVars: []string{"MFER_SEED"},
 					},
 				),