diff --git a/COMMANDS.md b/COMMANDS.md index 0b9d7cf..ad0ad91 100644 --- a/COMMANDS.md +++ b/COMMANDS.md @@ -151,6 +151,11 @@ fable-agent plinius godmode "improve explanation quality" - `benchmark run` - `-b, --benchmark ` +- `benchmark agents `: run the same prompt against built-in or JSON-defined harnesses + - `--with ` + - `--timeout-ms ` + - `--harnesses ` + - `--out ` - `benchmark duel `: record a two-implementer eval winner - `--a