You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Replaced the snapshot-style todowrite/todoread pair with one phased, op-based todo tool; the todowrite builtin extension id remains for loader compatibility.
Added
Added a /todo command for user-side todo management: view, overlay edit (Markdown checklist), copy, export/import (default TODO.md), and fuzzy-matched append/start/done/drop/rm; user edits persist as senpi.todo-state entries and notify the agent next turn.
Added antml as a valid compat.toolCallFormat in custom models.json provider and model definitions: ANTML <function_calls>/<invoke> text-tool protocol with Claude-Code-style failure tolerance.
Added built-in llama.cpp router support with /login connection setup and /llama Hugging Face model search and downloads, explicit loading, unloading, and live progress. See llama.cpp.
Added extension registration for complete pi-ai providers, including native authentication, model refresh, filtering, and streaming behavior.
Changed
Fixed
Fixed prompt-template defaults for all arguments (${@:-default} and ${ARGUMENTS:-default}) (#6695).
Fixed obsolete custom UI, custom tool, and custom editor examples in the extension documentation (#6735).
Fixed Kimi Coding sessions to show API-equivalent implied costs with the subscription indicator.
Fixed OpenAI Responses early stream endings to trigger automatic retry instead of ending the agent run (#6727).