Scripting Administration#

Config file#

config/simple_bytestorage_scripting-common.toml, section [scripting]. The server's values are what counts.

Key	Default	Range	Meaning
`enabled`	`true`		Master switch. `false` stops every Logic Bus (the blocks stay in the world).
`opsOnly`	`false`		`true` = only operators (permission level 2+) and the owner of a single-player/LAN world can save, run, stop or change the interval. Everyone else gets a read-only terminal.
`allowCrafting`	`true`		Allows `net.craft`.
`allowControl`	`true`		Allows scripts to change Controller settings: order priority/pause/cancel (`net.setPriority`, `net.pause`, `net.resume`, `net.cancel`), stock rules (`stock.set/clear`) and recipe settings (`recipes.*` changes). Reading is always allowed.
`allowAlerts`	`true`		Allows `alert(text)` (chat warnings to owners of bound Wireless Terminals).
`maxScriptLength`	`8000`	100 – 30000	Characters per script.
`minInterval`	`10`	1 – 1200	Minimum ticks between runs.
`maxInstructionsPerRun`	`50000`	1000 – 5,000,000	JS instruction budget per run.
`maxMillisPerRun`	`20`	1 – 500	Wall-clock budget per run.
`maxItemsPerRun`	`256`	1 – 100,000	Items a run may move.
`maxApiCallsPerRun`	`200`	10 – 10,000	API calls per run.

Limits#

All "per run" values apply to each run of tick() of each bus. A violation stops that bus with an error shown in the terminal. Buses run on the server thread, so keep maxMillisPerRun small on busy servers; total cost scales with the number of running buses.

Security model#

What the sandbox does:

Runs in Rhino's interpreter mode.
No Java access: a deny-all class filter is set, no Packages/java globals, and the API returns only plain JS values.
No file, network, or command access.
Limits for instructions, time, items, API calls, interpreter stack depth, and craft orders.
A script can only look up buses of its own network (side(dir, id), buses(id)); it cannot reach other networks.
Identifiers are sanitised on the server (letters, digits, _ - ., max 32).
Budget errors derive from Error, so scripts cannot swallow them with try/catch.
Known memory-amplifying built-ins (String.prototype.repeat, padStart, padEnd, Array.prototype.fill) are removed.

What it does not do:

There is no memory limit. A malicious script can still allocate a lot of memory with ordinary code (for example by repeatedly doubling a string or growing an array) within the instruction budget.
Any player with access to a Logic Terminal can read scripts, even in read-only mode.

Recommendation for public servers: set opsOnly = true, or don't install the addon at all.

Permissions summary#

Action	`opsOnly = false`	`opsOnly = true`
Read scripts and logs	anyone using the terminal	anyone using the terminal
Save / run / stop / interval (terminal) and assign IDs (Logic Bus GUI)	anyone	operators (level 2+) and the single-player owner

The server also verifies that the player is within ~12 blocks of the terminal and that the bus belongs to the terminal's network.

Not installing / removing the addon#

Without the addon jar, the blocks and items simply don't exist, and nothing else changes.
If you remove the addon from a world that already has Logic Buses, Forge will warn about missing registry entries and those blocks disappear. Remove them first if you care about the contents.
Clients need the addon too (it adds blocks and a GUI); the network channel versions must match (the addon's network protocol changes between versions, so a client with a different addon version can't join).

Found a mistake? This page is generated from docs/wiki/Scripting-Administration.md in the mod repository.