Thumbnail and Title Review Habit: a weekly compare-and-log routine

What you end up with

A short log of which packaging changes held on channels you watch, and two or three briefs for your own videos you approved, in about 30 minutes a week.

What you get

• The routine: monthly watchlist, daily capture, weekly compare and log, then briefs
• review_log.py: 11 error checks and 3 warnings, plus a weekly report with change labels and views per subscriber
• The one rule on copying, enforced: titles that share 4 or more words in a row with your log are stopped
• Watchlist, capture, brief and settings templates, and a fictional example with three channels
• AGENTS.md rules for Grok Bot, Claude Code, Codex and Hermes, and 15 tests

Who it is for

Anyone who wants to learn from title and thumbnail changes without copying anyone.

Works with

Solid chips: the agent is named in this listing's own tools list. Dashed chip: One Python script, CSV templates, a fictional example and AGENTS.md rules. Tested on Linux.

Tools needed

Cost to run: Free. Python is free.
Time to result: First weekly report in about 15 minutes

How it works

1. Run the example

   Unpack the file and run python3 review_log.py report on the example folder.

   Expected: A report with one confirmed survivor, one likely test and a ranked list.

2. Pick the watchlist

   Add up to 10 public channels to watchlist.csv and approve each with your name and a revisit date.

   Expected: check shows 0 errors.

3. Capture through the week

   Once or twice a day, record title, thumbnail note, views and subscribers for recent videos, by hand or from an official surface.

   Expected: snapshots.csv grows; check stays clean.

4. Report and brief

   Run the report on your review day, draft briefs in your own words, and approve or kill each one.

   Expected: Approved briefs that pass the copy check.

Sample prompts, free to use

Follow AGENTS.md. Run review_log.py report on this folder and summarize the survivors and over-performers in five lines. Then draft up to three briefs in briefs.csv as status draft, in my words, never reusing a logged title's wording.

Free preview of the file

# Thumbnail and Title Review Habit

**A weekly routine for learning from public packaging changes on channels you watch: compare what changed, log it with honest labels, and turn the angles that held into briefs for your own videos. In your words, with your yes.**

Version 1.0, 2026-11-03. By PaperChaseWebb Inc. Free. Works with Grok Bot, Claude Code, Codex and Hermes, or by hand. Python 3.8 or newer, no installs.

## Why this habit

Creators often change a video's title or thumbnail after it goes up. Some changes stick, some get rolled back. Watching which ones stick on channels close to yours is one of the cheapest ways to learn what your audience responds to. The trap is doing it badly: copying a title word for word, treating one change as proof, or pointing a scraper at a site without checking its rules.

This routine keeps the useful part and closes those gaps.

## The routine

**Once a month (10 minutes): pick the watchlist.** Up to 10 public channels that share your audience or format. You approve each one with your name and a revisit date about a month out. Channels without your approval are ignored by the checker.

**Once or twice a day (2 minutes): capture.** For new and recent videos on those channels, record the time, the title, the thumbnail (save the image, or write a short description of it), views and subscribers. Do it by hand or from an official surface. A scraper or browser automation needs its own approval from you after you have read the site's terms; it 

Worth knowing

• Labels describe what changed in public, not why. From the outside nobody can see a creator's numbers, so the log never calls a change a proven test.
• Capture by hand or from an official surface. Scrapers and browser automation need your own approval after you read the site's terms; they are not part of this routine.
• Views per subscriber is a rough signal. Small channels swing a lot.
• On 2026-11-03: 15 tests passed on Python 3.8, 3.9, 3.11, 3.12 and 3.13 on Linux. Agent rules guide; they were not loaded inside each harness for this release.

Details

Version 1.0 · Updated Nov 3, 2026 · Markdown version of this page
License: Personal and commercial use. Download terms

Related

Research report$39

Algorithm cheat sheets: what X, Instagram, YouTube and TikTok actually say they rank

by PaperChaseWebb

The ranking signals each platform has published, labeled Published, Via secondary or Lore, with the X For You weights read from its public source code and a checker that tells you when they change.

Works withClaude CodeCodexGrok BotHermesPlain .md report; the optional tools need Python 3.8 or newer

Solo business · Beginner · 10 min setup

Prompt pack$29

Content System Kit: run a content engine with AI agents and keep a human on the publish button

by PaperChaseWebb

A prompt pack and method for founders, small teams and agencies. Audience filters, one owned format, a 90-post test with review rules set in advance, a weekly batch loop with one approval gate, an editor queue with a rights check, transcript-first cleanup, a packaging watch on public data only, teardown without replication, a craft seat map, a cadence plan, an engagement honesty audit, a failure table, a worked example, a rules block and 14 prompts. Includes content_check.py, AGENTS.md with six owner stops, two skills and 15 templates.

Works withClaude CodeCodexGrok BotHermesPlain .md; the checker needs Python 3.8 or newer

Creative and video · Beginner · 1 hr setup