loadlisted

## Reading Task **User input**: $ARGUMENTS ### Process 1. **Resolve file references first**: - If inputs are already valid absolute/relative paths, use them directly - If inputs are ambiguous (partial names, no extension, shorthand), resolve them: - Use project context (CLAUDE.md, known directory structure) - Use Glob to find matching files if needed - Apply common sense (e.g., "index" in a docs project → `docs/index.md`) - Confirm the resolved paths exist before proceeding 2. **Run the analysis script** with resolved paths: ```bash ~/.claude/bin/read-files path1 path2 ... ``` 3. **Follow the output instructions exactly**: - `READ FULL: path` → Use Read tool on the file with no offset/limit - `CHUNK: path` → Use Read tool with the specified offset and limit for each chunk 4. **After reading**: Wait for further instructions (do not summarize or take additional actions)