Forge Code Files

Changes

🚀 Features

• feat: add forge zsh doctor command @tusharmath (#2113)
• feat: add IO Intelligence provider support @eminaruk (#2108)
• feat: add conversation delete command @dariuszkowalski-com (#2077)
• fix(mcp): show full error details in mcp show command @tusharmath (#2090)

📝 Documentation

• fix(zsh): update doctor instructions to use $FORGE_BIN variable @tusharmath (#2130)
• feat: add forge zsh doctor command @tusharmath (#2113)

🐛 Bug Fixes

• fix(zsh): update doctor instructions to use $FORGE_BIN variable @tusharmath (#2130)
• fix(conversation): iterate all context messages for usage calculation @tusharmath (#2131)
• fix(zsh): prepend forge rprompt existing RPROMPT instead of replacing @tusharmath (#2128)
• fix(zsh): add buffer redisplay before prompt reset in bracketed paste handler @amitksingh1490 (#2121)
• refactor(http): add http_ prefix to HttpInfra trait methods @tusharmath (#2119)
• fix: enable parallel tool calls by default @amitksingh1490 (#2111)
• fix(gpt-5.2): clarify patch tool usage for multi-file modifications @amitksingh1490 (#2112)
• fix(ui): prevent spinner cycling causing early exit in model selection @amitksingh1490 (#2086)
• fix(create-plan): enhance research and plan creation guideline @amitksingh1490 (#2109)
• fix(zsh): filter provider list to show only llm providers @amitksingh1490 (#2098)
• fix(mcp): show full error details in mcp show command @tusharmath (#2090)
• fix(openai): add content filter handling for GitHub Copilot responses @amitksingh1490 (#2096)

🚀 Performance

• perf(ui): make forge info 14x faster @tusharmath (#2123)

🧰 Maintenance

• chore(proto): match client with server proto defs @amitksingh1490 (#2120)
• refactor(http): add http_ prefix to HttpInfra trait methods @tusharmath (#2119)
• fix: error mesage encoding @tusharmath (#2118)
• refactor: move usage information from conversation to context message @tusharmath (#2101)
• refactor(orch): remove async compaction and execute before chat request @tusharmath (#2100)
• refactor(tools): add conversation id to llm tool output @tusharmath (#2099)
• refactor(display): remove unavailable marker and use empty marker in /provider @amitksingh1490 (#2097)