Skip to content

Fix completion policy tool batch mediation #316

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
chubes4 wants to merge 1 commit into
base: main
Choose a base branch
from
Open

Fix completion policy tool batch mediation #316

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
23 changes: 13 additions & 10 deletions src/Runtime/class-wp-agent-conversation-loop.php
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -541,6 +541,7 @@ private static function mediate_tool_calls(
$events = array();
$spin_signatures = array();
$complete = false;
$completion_stop_recorded = false;
$exceeded_budget = null;
$approval_required = null;
$runtime_tool_pending = null;
Expand Down Expand Up @@ -875,16 +876,18 @@ private static function mediate_tool_calls(

if ( $decision->isComplete() ) {
$complete = true;
$events[] = array(
'type' => 'completion_policy_stop',
'metadata' => array(
'tool_name' => $tool_name,
'turn' => $turn,
'message' => $decision->message(),
'context' => $decision->context(),
),
);
break;
if ( ! $completion_stop_recorded ) {
$events[] = array(
'type' => 'completion_policy_stop',
'metadata' => array(
'tool_name' => $tool_name,
'turn' => $turn,
'message' => $decision->message(),
'context' => $decision->context(),
),
);
$completion_stop_recorded = true;
}
}

$continuation = self::completion_policy_continuation( $decision, $tool_name, $turn );
Expand Down
45 changes: 45 additions & 0 deletions tests/conversation-loop-completion-policy-smoke.php
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -131,6 +131,51 @@ static function ( array $event ): bool {
agents_api_smoke_assert_equals( 'finish tool called', $stop_events[0]['metadata']['message'] ?? '', 'stop event carries decision message', $failures, $passes );
agents_api_smoke_assert_equals( array( 'tool_name' => 'client/finish', 'turn' => 3 ), $stop_events[0]['metadata']['context'] ?? array(), 'stop event carries decision context', $failures, $passes );

echo "\n[1a] Completion policy does not truncate same-turn tool batches:\n";
$policy_log = array();

$batch_result = AgentsAPI\AI\WP_Agent_Conversation_Loop::run(
array( array( 'role' => 'user', 'content' => 'write files' ) ),
static function ( array $messages ): array {
return array(
'messages' => $messages,
'tool_calls' => array(
array( 'name' => 'client/finish', 'parameters' => array( 'path' => 'index.html' ) ),
array( 'name' => 'client/work', 'parameters' => array( 'path' => 'style.css' ) ),
),
);
},
array(
'max_turns' => 5,
'tool_executor' => $executor,
'tool_declarations' => $tools,
'completion_policy' => $policy,
'should_continue' => static function (): bool {
return true;
},
)
);

$batch_tool_names = array_map(
static function ( array $tool_result ): string {
return (string) ( $tool_result['tool_name'] ?? '' );
},
$batch_result['tool_execution_results']
);

$batch_stop_events = array_values(
array_filter(
$batch_result['events'],
static function ( array $event ): bool {
return 'completion_policy_stop' === ( $event['type'] ?? '' );
}
)
);

agents_api_smoke_assert_equals( array( 'client/finish', 'client/work' ), $batch_tool_names, 'same-turn tool batch executes after completion policy marks complete', $failures, $passes );
agents_api_smoke_assert_equals( 1, count( $batch_stop_events ), 'same-turn batch records one completion stop event', $failures, $passes );
agents_api_smoke_assert_equals( 2, count( $policy_log ), 'policy was consulted for each same-turn tool result', $failures, $passes );

echo "\n[2] Completion policy coexists with should_continue — policy takes precedence:\n";
$policy_log = array();
$continue_called = 0;
Expand Down
Loading