towry · towry · Oct 23, 2025 · Oct 21, 2025 · Oct 22, 2025 · Oct 22, 2025
diff --git a/conf/bash/scripts/tmux-project-switcher.sh b/conf/bash/scripts/tmux-project-switcher.sh
@@ -88,7 +88,7 @@ if [[ -z $SELECTED_PATH ]]; then
 
     # Use binaries from PATH (assumes they're available)
     selected=$(zoxide query -l --exclude "$PWD" | \
-        path-git-format --filter --no-bare -f"{path} [{branch}]" | \
+        agpod vcs-path-info --filter --no-bare --format "{path} [{branch}]" | \
         awk -v home="$HOME" '{gsub(home, "~"); print}' | \
         fzf "${FZF_OPTS[@]}" \
             --preview "echo {} | awk -F '[' '{print \$1}' | awk -v home=\"\$HOME\" '{sub(/^~/,home)};1' | xargs -I % eza --color=always --icons=auto --group-directories-first --git --no-user --no-quotes --git-repos %" | \

diff --git a/conf/ghostty/config b/conf/ghostty/config
@@ -18,19 +18,19 @@ cursor-style = block
 cursor-invert-fg-bg = true
 selection-invert-fg-bg = true
 macos-option-as-alt = true
-# macos-titlebar-style = tabs
-macos-titlebar-style = transparent
-macos-window-shadow  = true
+macos-titlebar-style = tabs
+# macos-titlebar-style = transparent
+macos-window-shadow  = false
 adjust-underline-position = 15%
 adjust-underline-thickness = -40%
 adjust-box-thickness = -40%
-adjust-cell-height = 5%
+adjust-cell-height = 10%
 mouse-hide-while-typing = true
 focus-follows-mouse = false
 unfocused-split-opacity = 0.9
 window-padding-balance = true
 window-padding-color = background
-window-padding-x = 14
+window-padding-x = 8
 window-padding-y = 0
 window-save-state = default
 

diff --git a/conf/llm/docs/coding-rules.md b/conf/llm/docs/coding-rules.md
@@ -74,7 +74,8 @@
 - Verify that no existing debug or development processes are running.
 - Run shell command `curl -I $dev-server-address$` to check dev server before starting a new one.
 - Ask user to commit current changes before running lint/format to avoid unexpected diffs.
-- Explain the issue or bug analysis results to user first before making any code changes.
+- Explain the issue or bug analysis results to user first, then ask for confirmation to make any code changes.
+- When user frustrated, or issue is complex, use said "not working", "still broken", you need to try ask for advanced subagent for help.
 
 ## Tool Usage
 
@@ -88,6 +89,10 @@
 - **search web and scrape html**: Use the `brightdata` mcp tool to fetch the latest context from the web, like version, framework tools, and documentation.
 - *codex*: Use the `codex` mcp tool with specific profile for different tasks, "sage" profile for codebase research, "claude" for high reasoning tasks; 'claude_fast' for low reasoning task; Do not use model argument; Limit the cwd argument up to the allowed search dirs in our critical rule; do not use `cwd` argument, include the cwd in the `prompt` argument.
 
+#### playwright mcp tool
+
+- On first start, always double check if the page is loaded, the first load usally is a blank page initiated by the browser, you may need to navigate to the url again to ensure page is loaded.
+
 ### codex codebase research usage example
 
 ```json

diff --git a/conf/llm/opencode/agent/eng.md b/conf/llm/opencode/agent/eng.md
diff --git a/conf/llm/opencode/agent/oracle.md b/conf/llm/opencode/agent/oracle.md
@@ -1,5 +1,5 @@
 ---
-description: >
+description: >-
   Expert advisor for complex technical decisions requiring deep reasoning.
 
   When to use: (1) Complex debugging with unclear root cause; (2) Code review
@@ -8,15 +8,23 @@ description: >
   compatibility/performance constraints.
 
   When NOT to use: (1) Simple edits or quick fixes; (2) Command execution
-  (oracle cannot run shell); (3) Basic grep/read tasks.
+  (oracle cannot run shell); (3) Basic grep/read tasks. (4) Files searching or codebase research.
 
   How to use: Provide focused context - for code review: diff + intent +
   constraints; for debugging: logs + current behavior + what was tried + expected behavior; for
   refactoring: code snippets + test coverage + compatibility requirements.
+
+  Output: Summary, options with pros/cons, recommendation with rationale,
+  next steps, risks/assumptions.
+
+  Rules to use oracle: 
+  - Oracle is slower and costlier - keep scope tight, provide only necessary artifacts, tell oracle if it need more context, ask for it. And the caller should response for oracle's further context request.
+  - Do not ask for codebase details.
+  - Do not provide subjective analysis on the issue
 mode: subagent
-model: copilot/gpt-5
+model: "github-copilot/gpt-5"
 reasoningSummary: concise
-textVerbosity: low
+textVerbosity: middle
 tools:
   write: false
   edit: false
@@ -44,7 +52,7 @@ You are the Oracle - an expert AI advisor for complex technical decisions.
 
 # Core Principles
 
-- Verify correctness with provided context
+- Verify correctness with provided context, ignore the subjective analysis the user provided.
 - Prioritize project conventions over general best practices
 - *Maintainability*: Long-term convenience over short-term hacks
 - Avoid over-engineering and unnecessary complexity