v1.0.0 - Major Architecture Rewrite

[Tarquinen]

Summary

Major rewrite to use OpenCode's native hooks API, removing provider-specific fetch wrapper code and dramatically simplifying the architecture.

Highlights

• Complete architecture rewrite using experimental.chat.messages.transform and experimental.chat.system.transform hooks
• Removed ~1,400 lines of provider-specific format handlers (Anthropic, Bedrock, Gemini, OpenAI)
• New features: System prompt transformation, AI nudge system, custom config path support
• Prune tool enabled by default for more aggressive context management

What's Changed

Architecture

• Removed lib/fetch-wrapper/ directory with 6+ format adapters
• Simplified state management, removed ID mapping layer
• Consolidated tool cache logic

New Features

• System prompt transformation hook for synthetic context injection
• AI nudge system reminds LLM to use prune tool after N tool results
• Support for $OPENCODE_CONFIG_DIR config location

Bug Fixes

• Skip inserting prunable tools list when empty
• Properly reset nudge counter after prune tool use
• Await session state load before syncing tool cache
• Prevent compacted tools from appearing in prunable list

Breaking Changes

• On-idle strategy now marked as legacy and disabled by default

Full Changelog: v0.4.17...v1.0.0