diff --git a/docs/TODO.md b/docs/TODO.md
index bf5577b..30bc410 100644
--- a/docs/TODO.md
+++ b/docs/TODO.md
@@ -1,115 +1,7 @@
-# Project Tasks
-
-<!--
-Purpose:
-- This document is the user-visible task list and agent-visible project queue.
-
-Instructions for AI agents:
-- Do not add tasks to the `## User tasks` section.
-- Do add tasks to the `## Agent tasks` section. Include all open work from agent-managed handoff documents.
-- Use `- [ ]` to indicate open work and `- [x]` for work completed during the current session.
-- Remove completed standalone agent tasks after recording their outcomes in `docs/STATUS.md`.
--->
+# TODO.md
 
 ## User tasks
 
-- [ ] Explore bounded Go utilities, including a worker-pool utility where justified. The docmend product remains Python:
-
-````markdown
-For a standalone utility that processes many files efficiently, use a **Worker Pool** pattern. Do not treat this example as a docmend product-migration design. Instead of spawning one million goroutines at once (which could overwhelm OS file descriptors or memory), spawn a fixed number of "workers" (e.g., 100) that constantly pull file paths from a queue until the queue is empty.
-
-Here are the three core pieces that make this work:
-
-1. **The Jobs Channel (`chan string`):** This acts as a thread-safe queue. You push file paths into it, and the workers pull paths out of it.
-2. **The WaitGroup (`sync.WaitGroup`):** This acts as a counter. It tells the main program to pause and wait until every single worker has finished its job. Without this, your program would exit immediately before the workers get a chance to run.
-3. **The Workers:** These are lightweight goroutines running an infinite loop. They pull a file path from the channel, process it, and go back for another one. When the channel is closed, the loop automatically breaks.
-
-Here is a complete, runnable example of this pattern:
-
-```go
-package main
-
-import (
-	"fmt"
-	"math/rand"
-	"sync"
-	"time"
-)
-
-// 1. The Worker Function
-// It takes the channel of file paths and the WaitGroup pointer
-func fileProcessor(workerID int, jobs <-chan string, wg *sync.WaitGroup) {
-	// Ensure we tell the WaitGroup we are done when this function exits
-	defer wg.Done()
-
-	// This loop pulls from the channel until the channel is closed
-	for filePath := range jobs {
-		fmt.Printf("Worker %d picked up: %s\n", workerID, filePath)
-
-		// ---------------------------------------------------------
-		// YOUR LOGIC GOES HERE:
-		// 1. Open the file (os.Open)
-		// 2. Parse it (bufio.Scanner, goquery, or goldmark)
-		// 3. Write the manipulated data to a new file
-		// ---------------------------------------------------------
-
-		// Simulating the time it takes to read/write a file
-		time.Sleep(time.Millisecond * time.Duration(rand.Intn(300)))
-
-		fmt.Printf("Worker %d finished: %s\n", workerID, filePath)
-	}
-}
-
-func main() {
-	// A mock list of files you might get from scanning a directory using filepath.WalkDir
-	filesToProcess := []string{
-		"index.html", "about.html", "readme.md",
-		"data1.txt", "data2.txt", "contact.html",
-	}
-
-	// Create a buffered channel to hold our file paths
-	jobs := make(chan string, len(filesToProcess))
-	var wg sync.WaitGroup
-
-	// Determine how many concurrent workers to run.
-	// For CPU/Disk heavy tasks, runtime.NumCPU() is a good starting point.
-	numWorkers := 3
-
-	fmt.Println("Starting workers...")
-
-	// 2. Spin up the workers
-	for w := 1; w <= numWorkers; w++ {
-		wg.Add(1) // Increment the WaitGroup counter for each worker we start
-		go fileProcessor(w, jobs, &wg)
-	}
-
-	// 3. Load up the queue
-	// We send all the file paths into the jobs channel
-	for _, file := range filesToProcess {
-		jobs <- file
-	}
-
-	// Close the channel. This tells the workers "no more files are coming."
-	// It does NOT destroy the channel; workers will finish pulling remaining items.
-	close(jobs)
-
-	// 4. Wait for completion
-	// The main function blocks here until wg.Done() is called enough times
-	// to bring the WaitGroup counter back down to zero.
-	wg.Wait()
-
-	fmt.Println("All files processed successfully!")
-}
-
-```
-
-### Pro-Tips for your specific use case:
-
-- **How many workers?** If you are reading/writing from an SSD, try `100` to `500` workers. If you push it too high, your OS will complain about having too many open files.
-- **Handling Errors:** In a real-world scenario, you will want a second channel (e.g., `results chan error`) so workers can send back errors if a file fails to read or write. Your main function can then log those errors without stopping the rest of the pool.
-- **Finding the files:** You will want to use Go's `filepath.WalkDir` in the main function to rapidly scan your directory structure and feed those found paths directly into the `jobs` channel.
-````
-
 ## Agent tasks
 
 - [ ] Support the owner's first staged real-library rollout under spec section 18.4.