Settings by goal #
On this page
Maple Harness aims at three things at once: to be efficient (fewer tokens for the same work), efficacious (the work stays as good, and defaults are safe) and effective (whole tasks succeed and waste stops early). This page maps a goal to the settings and commands that serve it. It makes no promise about your bill: results vary from run to run and task to task, so measure your own, and turn off anything that does not help on your work. Every setting listed has an off switch.
In plain words #
- What it is. A map from what you want (spend less, keep quality, stop runaways, get reliable results) to the setting that serves it.
- What to do. Find your goal below, change the setting where it says (most are on the setup page's Models & routing tab, or on Advanced, which Find a setting reaches by name), and measure the result on your own work.
I want to spend less on long runs #
| Do this | Where | Trade-off |
|---|---|---|
| Keep one model for a long task instead of switching | /model <name>, or /model auto (it keeps your model when it fits) | A fixed model may be more than a small job needs |
| See where the money goes | /cost, the setup page's Cache card, maple-proxy cache | None |
| Keep the provider's cache working | providers[].prompt_cache (on by default), prefix_stabilizer (on by default) | None for the defaults |
| Trim old tool output in very long conversations | context_trim on the Guards card (off by default) | The model may ask for trimmed output again |
| Keep noisy searches out of the main conversation | @name <task> or /agent, with a subagent | A subagent starts with only the task you give it |
| Stop a stuck run early | loop_guard.mode (on by default) | It may end a run you would have let continue |
| Shorten long failing logs | tool_output.failed_max_lines (on by default) | The model sees less of a long log |
| Cap what a provider can cost per day | budgets[] on the Budgets card | Requests are refused, or fall back, when reached |
| Start fresh tasks in fresh sessions | /clear, or a new session | You lose the earlier context |
I want the work to stay as good #
| Do this | Where | Trade-off |
|---|---|---|
| Let the compiler check the model's edits | lsp (on by default), /lsp to see what is running | The first edit starts the server in the background |
| Keep the safe defaults | permission.mode: ask, sandbox.mode, file_policy.* | More questions, fewer surprises |
| Prefer the tools that show a diff | bash_file_writes: warn (default) or deny | deny refuses shell writes |
| Plan before changing anything | /plan | An extra step |
| Use a capable model for hard work | /model, auto_model_map | Costs more per turn |
| Keep automatic choice away from models that decline long tasks | auto_avoid | A pattern may hide a model you would have liked |
| Check your setup | maple-harness doctor | None |
I want to stop runaways #
| Do this | Where | Trade-off |
|---|---|---|
| End a turn that keeps failing the same way, with a summary | loop_guard.mode: stop, loop_guard.stop_after | See above |
| Cut a reply that repeats itself | degenerate_guard (the repeat guard) on the proxy's Guards card | A repetitive answer could be cut short |
| Refuse or trim a request bigger than the model's window | context_guard | truncate forgets old turns |
| Cap the turns a subagent may take | subagents.max_turns | A long job may need more |
| Limit how fast a client may call the proxy | server.rate_limit.* | Bursts are slowed |
| Move to another model when one runs out | model_fallbacks | The fallback may be different or weaker |
I want results I can rely on #
| Do this | Where | Trade-off |
|---|---|---|
| Get told when a session needs you | notifications.*, maple-harness notify test | None |
| Resume where you left off | --continue, --resume, /sessions | Old sessions take disk; cleanup removes them |
| Keep a checklist for a big request | ledger: auto, /ledger | One extra model call |
| Make the model carry on when it stops early | on by default; disable_auto_continue turns it off | None for the default |
| Check a change against your own tasks | maple-harness --prompt with --json-summary, repeated with a setting on and off | It costs the runs you make |
Read next: Proxy settings, Commands and keys and Harness settings.