@@ -9,7 +9,7 @@ A TUI from the future. We want feature parity with a GUI editor: mouse selection
99 ## Minimum viable product:
1010
1111 - Agentic UI:
12- - Support for highlighted code snippets. Git diffs and shell outputs. (Mini editor components?)
12+ - Support for clicking file references/diffs/snippets to open in Editor side.
1313 - Optional Human-in-the-middle LLM tool call loop (with configurable permissions in global config)
1414 - Auto Context, LLM is aware of file/project in question probably via tree-sitter?
1515 - AGENTS.md support
@@ -23,7 +23,7 @@ A TUI from the future. We want feature parity with a GUI editor: mouse selection
2323
2424 - LLM Tools:
2525 - Run in sandboxed version of the working directory to allow multiple agents working on the same file. Use git to merge sandboxed changes to real repo.
26- - Grep, Read, Write: optimized for code changes (LSP and linting checks after each edit). These appear in agentic UI as a diff. Clicking on it takes us to that edit in the editor tab.
26+ - Grep, Read, Write: optimized for code changes (LSP and linting checks after each edit). These appear in agentic UI as a diff message (not actual diff). Clicking on it takes us to that edit in the editor tab.
2727 - Undo: Either perform an intelligent step back or use git to reset.
2828 - Subagent tool: Performs a single task based on prompt.
2929 - Mutation tools create a new branch worktress for the sandboxed environment. Perform the mutation. Once the user approves the mutation or tells the agent to merge, a git merge happens.
@@ -71,7 +71,7 @@ A TUI from the future. We want feature parity with a GUI editor: mouse selection
7171
7272 - True ELM architecture (no exceptions ever).
7373 - Strict project structure with separation of concern by internal modules.
74-- Go (see https://github.com/tj/go-tea) For ELM architecture in TUI app.
74+- Go (see https://github.com/tj/go-tea) For ELM architecture in TUI app (documentation).
7575 - BubbleTea:
7676 - Base library for TUI, use their bubbles, and all of the ecosystem.
7777 - Golden files based testing for the TUI.
@@ -108,4 +108,22 @@ State Management. You have three asynchronous "gods" to serve:
108108 3. LLM Agent (Thinking/Tools) -> Slow (seconds).
109109
110110 BubbleTea's ELM architecture (Update() loop) is perfect for this if you keep the heavy lifting (LLM/LSP) in separate Go routines that send messages back to the main thread. If you block the main thread, the UI will freeze.
111-Confidence Level: High for a prototype. The ecosystem is ready.
111+Confidence Level: High for a prototype. The ecosystem is ready.
112+
113+## MVP Phases:
114+
115+1. Phase 1: Basic TUI with editor + agent pane (no LSP, no git sandbox)
116+ - File viewer with syntax highlighting
117+ - Agent conversation with tool calls (Read/Write/Grep)
118+ - Simple diff preview before applying
119+2. Phase 2: Git integration
120+ - Show git status in editor
121+ - Ctrl+g for diffs
122+ - Agent commits to real repo (no sandbox yet)
123+3. Phase 3: LSP integration
124+ - Start with diagnostics only
125+ - Add go-to-definition later
126+4. Phase 4: Advanced features
127+ - Sandboxed worktrees
128+ - Sub-agents
129+ - Context optimization with tree-sitter