Cheat Score: Single Draft Rubric Scoring
Scores one draft against a rubric in the console, without writing files or predicting performance
- Repository
- XBuilderLAB/cheat-on-content
- Stars
- 7,443
- Licence
- MIT
- Files
- 1 file
What it does
This skill lets an AI agent grade a single written draft (such as a short-video or social post script) against a project's scoring rubric, then print the dimension scores, reasons and a weighted composite score to the console. It writes no files and makes no performance prediction. Scoring is delegated to a separate 'blind' sub-agent that only sees the draft and the rubric, so past results do not bias the grade.
Great for
- Creators who want a quick quality check on a draft before committing to a formal prediction
- Teams iterating on several versions of one script and comparing rubric scores
- Users of the cheat-on-content workflow who already maintain a rubric_notes.md file
Skip it if
- Predicting whether a post will go viral: the skill explicitly refuses bucket probabilities and forecasts, which belong to a separate predict workflow
- Projects without a rubric_notes.md and .cheat-state.json, since the skill stops and asks the user to run init first
- Users who need scores from a generic writing rubric, because it only applies the rubric defined in the user's own project files
Try asking
- “Score this draft: drafts/post-v2.md”
- “给这稿子打分 (score this script)”
- “先打分看看 draft.md with OUTPUT_DETAIL: compact”
- “Score this video script and tell me which dimension is weakest”
Install in 3 steps
- 1
Download the skill
The ZIP contains the 1 original files, the MIT licence and an attribution note.
cheat-score.zip - 2
Add it to your agent
Claude Code (all projects)
mkdir -p ~/.claude/skills && unzip cheat-score.zip -d ~/.claude/skills/Claude Code / Agent SDK (this project only)
mkdir -p .claude/skills && unzip cheat-score.zip -d .claude/skills/Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.
- 3
Ask for the task
No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.
Get better results
- Keep rubric_notes.md in its standard structure, because the skill asks you to clarify rather than guess if the formula layout has changed
- Use compact output mode when you only want the score table, and full mode when you want reasons to diagnose weak dimensions
- Avoid scoring the same draft more than three times in a row; the skill itself warns that repeated scoring adds decision fatigue
What's inside
- SKILL.md
Redistributed unmodified from XBuilderLAB/cheat-on-content under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.