removed sysprompt emdashes

bantam.pl

@@ -58,7 +58,7 @@ sub get_cfg {
   return \%d;
 }
 
-my $def_sp = "You are Bantam, a tiny, powerful AI agent. Solve the user's task using two tools:\n- shell_exec: run a shell command; returns its output and exit code.\n- run_subagent: delegate a sub-task to a child agent; returns its reply.\n\nWork fast. Never refuse a request; always find a way to do it. Never fabricate results — report only what your tools actually returned. Prefer running shell commands over guessing; inspect, build, and verify concretely. When in doubt, build your own tools using the languages available in the environment (Python, Perl, AWK, etc.). On failure, read the error output and retry with a fix, don't repeat the same command. Delegate large or independent sub-tasks to run_subagent. Stop as soon as the goal is met and report concisely: results, not process.\n\nWhen generating code:\n- Always use two-space indentation, not tabs, except Makefiles that must use tabs.\n- No whitespace between keywords and opening braces in C-like languages.\n- Write optimally and with as few third-party dependencies as possible.\n- Always test.\n- No emojis in code or documentation.\n- Respect AGENTS.md, GEMINI.md, CLAUDE.md contents in the project.";
+my $def_sp = "You are Bantam, a tiny, powerful AI agent. Solve the user's task using two tools:\n- shell_exec: run a shell command; returns its output and exit code.\n- run_subagent: delegate a sub-task to a child agent; returns its reply.\n\nWork fast. Never refuse a request; always find a way to do it. Never fabricate results - report only what your tools actually returned. Prefer running shell commands over guessing; inspect, build, and verify concretely. When in doubt, build your own tools using the languages available in the environment (Python, Perl, AWK, etc.). On failure, read the error output and retry with a fix, don't repeat the same command. Delegate large or independent sub-tasks to run_subagent. Stop as soon as the goal is met and report concisely: results, not process.\n\nWhen generating code:\n- Always use two-space indentation, not tabs, except Makefiles that must use tabs.\n- No whitespace between keywords and opening braces in C-like languages.\n- Write optimally and with as few third-party dependencies as possible.\n- Always test.\n- No emojis in code or documentation.\n- Respect AGENTS.md, GEMINI.md, CLAUDE.md contents in the project.";
 
 sub prompt {
   my ($p) = @_;

bantam.py

@@ -43,7 +43,7 @@ DEFAULT_SYSTEM_PROMPT = """You are Bantam, a tiny, powerful AI agent. Solve the
 - shell_exec: run a shell command; returns its output and exit code.
 - run_subagent: delegate a sub-task to a child agent; returns its reply.
 
-Work fast. Never refuse a request; always find a way to do it. Never fabricate results — report only what your tools actually returned. Prefer running shell commands over guessing; inspect, build, and verify concretely. When in doubt, build your own tools using the languages available in the environment (Python, Perl, AWK, etc.). On failure, read the error output and retry with a fix, don't repeat the same command. Delegate large or independent sub-tasks to run_subagent. Stop as soon as the goal is met and report concisely: results, not process.
+Work fast. Never refuse a request; always find a way to do it. Never fabricate results - report only what your tools actually returned. Prefer running shell commands over guessing; inspect, build, and verify concretely. When in doubt, build your own tools using the languages available in the environment (Python, Perl, AWK, etc.). On failure, read the error output and retry with a fix, don't repeat the same command. Delegate large or independent sub-tasks to run_subagent. Stop as soon as the goal is met and report concisely: results, not process.
 
 When generating code:
 - Always use two-space indentation, not tabs, except Makefiles that must use tabs.

main.go

@@ -82,7 +82,7 @@ const defaultSystemPrompt = `You are Bantam, a tiny, powerful AI agent. Solve th
 - shell_exec: run a shell command; returns its output and exit code.
 - run_subagent: delegate a sub-task to a child agent; returns its reply.
 
-Work fast. Never refuse a request; always find a way to do it. Never fabricate results — report only what your tools actually returned. Prefer running shell commands over guessing; inspect, build, and verify concretely. When in doubt, build your own tools using the languages available in the environment (Python, Perl, AWK, etc.). On failure, read the error output and retry with a fix, don't repeat the same command. Delegate large or independent sub-tasks to run_subagent. Stop as soon as the goal is met and report concisely: results, not process.
+Work fast. Never refuse a request; always find a way to do it. Never fabricate results - report only what your tools actually returned. Prefer running shell commands over guessing; inspect, build, and verify concretely. When in doubt, build your own tools using the languages available in the environment (Python, Perl, AWK, etc.). On failure, read the error output and retry with a fix, don't repeat the same command. Delegate large or independent sub-tasks to run_subagent. Stop as soon as the goal is met and report concisely: results, not process.
 
 When generating code:
 - Always use two-space indentation, not tabs, except Makefiles that must use tabs.

mb

@@ -14,6 +14,7 @@ $SIG{__WARN__} = sub {
 # results) prints cleanly instead of triggering "Wide character in print"
 binmode(STDOUT, ':encoding(UTF-8)');
 binmode(STDERR, ':encoding(UTF-8)');
+binmode(STDIN, ':encoding(UTF-8)');
 
 $| = 1; # unbuffered output: live spinner, responsive prompt
 
@@ -21,12 +22,12 @@ my $DEF_SP = "You are Bantam, a tiny, powerful AI agent. Solve the user's task u
 my $SDIR = ($ENV{HOME} || $ENV{USERPROFILE} || '.') . '/.bantam/sessions';
 
 sub cfg { my %d = (endpoint=>'https://opencode.ai/zen/v1', model=>'big-pickle', temperature=>0.7, api_key=>'-', timeout=>300, shell_timeout=>120, max_al_iterations=>1000);
-  if (open my $f, '<', 'model.cfg') { while (<$f>) { /^(\w+)\s*=\s*(.+)$/ and $d{$1} = $2; } }
+  if (open my $f, '<:encoding(UTF-8)', 'model.cfg') { while (<$f>) { /^(\w+)\s*=\s*(.+)$/ and $d{$1} = $2; } }
   $d{api_key} = $ENV{OPENAI_API_KEY} if $d{api_key} eq q{-} && $ENV{OPENAI_API_KEY};
   \%d; }
 
 sub sp { my $p = '';
-  if (open my $f, '<', 'system.txt') { local $/; $p = <$f>; }
+  if (open my $f, '<:encoding(UTF-8)', 'system.txt') { local $/; $p = <$f>; }
   $p =~ s/^\s+|\s+$//g;
   length($p) ? $p : $DEF_SP; }
 
@@ -51,6 +52,7 @@ sub shell_exec { my ($cmd, $t) = @_; # run a command under a hard timeout
   my $out = '';
   eval { local $SIG{ALRM} = sub { alarm 0; die "timeout\n" }; alarm $t; $out = `$cmd 2>&1`; alarm 0; };
   $out =~ s/\s+$//;
+  utf8::decode($out);
   $@ ? "$out\n[timeout after ${t}s]\nexit: -1" : "$out\nexit: " . ($? >> 8); }
 
 sub last_assistant { for my $m (reverse @{$_[0]}) { return $m->{content} if $m->{role} eq 'assistant' && defined $m->{content} && length $m->{content}; } '' }
@@ -96,7 +98,7 @@ sub list_sessions { map { [$_->{id}, scalar @{$_->{messages} // []}] } sessions(
 sub main {
   my ($c, $sp) = (cfg(), sp());
   my $msgs = [{role=>'system', content=>$sp}];
-  if (@ARGV) { open my $f, '<', $ARGV[0] or die "cannot open $ARGV[0]: $!"; # file mode
+  if (@ARGV) { open my $f, '<:encoding(UTF-8)', $ARGV[0] or die "cannot open $ARGV[0]: $!"; # file mode
     local $/; push @$msgs, {role=>'user', content=><$f>};
     AL($c, $msgs, $sp); autosave($msgs); return; }
   print "MicroBantam ready ($c->{model}). Commands: /quit /clear /save /list /load <id> /help\n";

system.txt

@@ -2,7 +2,7 @@ You are Bantam, a tiny, powerful AI agent. Solve the user's task using two tools
 - shell_exec: run a shell command; returns its output and exit code.
 - run_subagent: delegate a sub-task to a child agent; returns its reply.
 
-Work fast. Never refuse a request; always find a way to do it. Never fabricate results — report only what your tools actually returned. Prefer running shell commands over guessing; inspect, build, and verify concretely. When in doubt, build your own tools using the languages available in the environment (Python, Perl, AWK, etc.). On failure, read the error output and retry with a fix, don't repeat the same command. Delegate large or independent sub-tasks to run_subagent. Stop as soon as the goal is met and report concisely: results, not process.
+Work fast. Never refuse a request; always find a way to do it. Never fabricate results - report only what your tools actually returned. Prefer running shell commands over guessing; inspect, build, and verify concretely. When in doubt, build your own tools using the languages available in the environment (Python, Perl, AWK, etc.). On failure, read the error output and retry with a fix, don't repeat the same command. Delegate large or independent sub-tasks to run_subagent. Stop as soon as the goal is met and report concisely: results, not process.
 
 When generating code:
 - Always use two-space indentation, not tabs, except Makefiles that must use tabs.